What Is a NASH Score and What Do the Results Mean?

A “NASH score” most commonly refers to the NAFLD Activity Score, or NAS, a number between 0 and 8 that pathologists assign after examining a tiny sample of liver tissue under a microscope. It grades three features of liver damage: fat accumulation, inflammation, and a specific type of cell injury called hepatocyte ballooning. But the term gets applied loosely, and in practice your doctor might be talking about any of several scoring tools used to assess fatty liver disease, including blood-based calculators and imaging results. Understanding which score you are looking at, and what it actually tells you about your liver health, matters because different scores answer different questions.

The NAS Itself and What Each Number Means

The NAS was developed by the NASH Clinical Research Network to standardize how pathologists describe what they see in a liver biopsy. It adds up three individual grades:

Steatosis (0–3): How much fat is present in liver cells, from none or minimal (0) to more than two-thirds of cells containing fat droplets (3).
Lobular inflammation (0–3): The number of inflammatory foci visible in the tissue, ranging from none (0) to more than four per field of view (3).
Hepatocyte ballooning (0–2): Whether liver cells show a characteristic swollen appearance, from none (0) to many (2).

The total NAS ranges from 0 to 8. A score of 5 or higher has often been used as a rough threshold suggesting steatohepatitis, the more aggressive form of fatty liver disease. But this threshold is less reliable than many people assume. In a large study examining the relationship between NAS values and a pathologist’s overall diagnosis, only about 75% of biopsies with a definitive diagnosis of steatohepatitis had a NAS of 5 or above. Meanwhile, roughly 29% of biopsies scoring 4 or below still showed steatohepatitis on expert review, and only 42% of those lower-scoring biopsies were categorized as definitively “not steatohepatitis.”¹ The score is useful for tracking change over time, but it was never designed to be a standalone diagnostic tool.

The Name Change You Will Encounter

If you have been reading about fatty liver disease recently, you may have noticed that the terms NAFLD and NASH are being phased out. A multisociety consensus in 2023 replaced NAFLD with “metabolic dysfunction-associated steatotic liver disease,” or MASLD, and the inflammatory form formerly called NASH is now called MASH (metabolic dysfunction-associated steatohepatitis). The old names were considered problematic because they relied on ruling out other causes of liver disease (“nonalcoholic”) and used language some patients found stigmatizing.² The underlying scoring systems have not changed, so a NAS of 5 means the same thing it always did. You will just see it discussed in the context of MASLD and MASH rather than NAFLD and NASH in newer medical records and research.

Blood-Based Fibrosis Scores Your Doctor Might Order

Most people who hear the phrase “NASH score” from their doctor are not actually getting a biopsy. They are far more likely receiving results from a blood-based calculator that estimates the degree of liver fibrosis, which is scarring. Two of the most common are the FIB-4 index and the NAFLD Fibrosis Score (NFS). These are not measuring inflammation or fat the way the NAS does. They are estimating how much structural damage has built up over time.

The FIB-4 index uses your age, platelet count, and two liver enzymes (AST and ALT) to sort you into risk categories. A score below 1.3 is generally considered low risk for advanced fibrosis, while a score of 3.25 or above signals high risk.³ Scores in between land in an indeterminate zone that usually triggers further testing. For people 65 and older, the cutoffs are adjusted upward because age itself raises the FIB-4, and using the standard thresholds would flag too many healthy older adults.⁴

The NAFLD Fibrosis Score works similarly but folds in more variables, including body mass index, blood sugar status, and albumin levels. Scores below about −1.5 suggest low probability of advanced fibrosis, while scores above 0.67 suggest high probability.⁵ As with FIB-4, there is a wide indeterminate zone where neither ruling in nor ruling out fibrosis is possible from blood tests alone.

A practical problem with these calculators is that they depend on having all the right lab values available at the same time. One real-world study of patients with type 2 diabetes and fatty liver disease found that only about 36% had all the labs needed to calculate FIB-4, with liver enzyme values being the most frequently missing piece.⁶ If your doctor has not ordered the full panel, these scores simply cannot be computed.

Specialized Blood Panels for Fibrosis

Beyond the simple calculators, there are commercial blood tests that measure proteins directly involved in the scarring process. The Enhanced Liver Fibrosis (ELF) test is one of the better-studied options. It measures three markers related to how the liver deposits and remodels scar tissue, then combines them into a single number. An ELF score of 9.8 or above has been used as a threshold suggesting advanced fibrosis, with moderate sensitivity and good specificity at that cutoff.⁷

In head-to-head comparisons, the ELF test and FIB-4 perform about equally well at detecting moderate-to-advanced fibrosis. Where the ELF test appears to pull ahead is in detecting cirrhosis specifically, where its accuracy has been measured as meaningfully higher than FIB-4.⁸ The ELF test is not widely available everywhere and costs more than a simple calculator, so it tends to be reserved for cases where initial screening leaves the picture unclear.

Imaging Scores and What They Measure

Transient elastography, commonly performed using a device called FibroScan, gives two numbers at once: a liver stiffness measurement (LSM) in kilopascals (kPa) and a controlled attenuation parameter (CAP) score that estimates fat content. These are the imaging-based “scores” most likely to come up in a fatty liver workup.

Liver stiffness correlates with fibrosis. A value below about 8 kPa is generally considered normal or indicating minimal fibrosis, while values above that suggest increasing degrees of scarring.⁹ A meta-analysis found that the accuracy of liver stiffness measurement was strong for ruling out or confirming cirrhosis and fairly good for intermediate fibrosis stages.¹⁰

The CAP score measures how much ultrasound is attenuated by fat in the liver. Higher values indicate more fat. CAP performs well at detecting whether any meaningful fat is present (steatosis grade 1 or above) and moderately well at distinguishing higher grades of fat accumulation.¹¹ The limitation is that CAP values can be less reliable in people with very high BMI or narrow intercostal spaces, situations where the ultrasound probe struggles to get a clean reading.

MRI-based methods provide even more detailed information. MRI proton density fat fraction (MRI-PDFF) is considered the most accurate non-invasive way to quantify liver fat, and MR elastography (MRE) measures liver stiffness with higher precision than ultrasound-based elastography. In clinical research, MRE-based liver stiffness has shown strong ability to identify patients with advanced fibrosis, while MRI-PDFF excels at identifying steatohepatitis itself.¹² These MRI tests are expensive and not available at every facility, so they tend to be reserved for research settings or cases where other tools have given ambiguous results.

The FAST Score Combines Multiple Measures

Recognizing that no single non-invasive test captures everything about fatty liver disease, researchers developed the FAST score, which stands for FibroScan-AST. It combines liver stiffness, CAP, and the blood enzyme AST into a single number designed to identify people who have both active steatohepatitis and significant fibrosis, the patients most likely to benefit from treatment.¹³

The FAST score runs from 0 to 1. A cutoff of 0.35 was chosen to have high sensitivity, meaning very few true cases are missed, while a cutoff of 0.67 was chosen for high specificity, meaning scores above that level are very likely to be real. In the original validation across international cohorts, the positive predictive value at the higher cutoff ranged widely depending on the population being tested, so a high FAST score is a strong signal but not a guarantee.¹³ The FAST score is increasingly used in screening algorithms and clinical trial enrollment to decide who needs a liver biopsy and who can be monitored non-invasively.

Why Fibrosis Stage Matters More Than the NAS for Your Long-Term Outlook

This is the single most important thing to understand about fatty liver scores: the feature that best predicts whether you will develop serious liver problems or die earlier is fibrosis stage, not the NAS. Researchers have confirmed this repeatedly. In one large prospective study, all-cause mortality rose steadily with fibrosis stage, from about 0.32 deaths per 100 person-years for stages 0 through 2, to 0.89 for stage 3 (bridging fibrosis), to 1.76 for stage 4 (cirrhosis).¹⁴

A systematic review and meta-analysis put numbers on the increased risk at each fibrosis stage compared to no fibrosis: roughly 1.6 times the mortality risk at stage 1, about 2.5 times at stage 2, about 3.5 times at stage 3, and over 6 times at stage 4.¹⁵ The risk of liver-specific death escalated even more dramatically, reaching extraordinarily high levels at advanced stages. A separate meta-analysis found a similar pattern, with statistically significant mortality increases beginning at stage 2.¹⁶

Critically, one study found that accounting for whether someone had steatohepatitis or not, on top of their fibrosis stage, did not meaningfully change the mortality estimates.¹⁷ In other words, two patients with the same amount of scarring face similar risks regardless of whether their biopsy shows active inflammation. This is why the clinical world has increasingly shifted its attention from asking “do you have NASH?” to asking “how much fibrosis do you have?”

The Limits of Liver Biopsy

Given that the NAS comes from a biopsy, it is worth knowing that biopsies are imperfect. A liver biopsy captures about one fifty-thousandth of the organ, and fatty liver disease does not distribute itself evenly. A study that took two separate biopsy samples from the same patients during the same procedure found that discordance of one fibrosis stage or more occurred in 41% of cases. Even more striking, about 35% of patients who had bridging fibrosis on one sample showed only mild or no fibrosis on the other.¹⁸

Hepatocyte ballooning, one of the three components of the NAS, was missed entirely in about 24% of patients when only a single biopsy was examined. The best negative predictive value a single biopsy could achieve for the diagnosis of steatohepatitis was only about 0.74, meaning roughly one in four patients told they did not have NASH actually did.¹⁸ Beyond accuracy, biopsies carry small but real risks including bleeding and pain, and they cannot be repeated casually during routine follow-up.¹⁹ This combination of sampling error and procedural risk is a major reason the field has invested so heavily in non-invasive alternatives.

How Screening Typically Works in Practice

If you have risk factors for fatty liver disease, such as type 2 diabetes, obesity, or metabolic syndrome, the most common screening pathway starts with a FIB-4 calculation from routine blood work. Proposed algorithms suggest checking FIB-4 annually in high-risk patients. Those with a low score can be reassured and monitored. Those with an indeterminate or high score are then referred for transient elastography, and if that also suggests advanced fibrosis, a specialist evaluation follows.²⁰ Biopsy is typically reserved for cases where the non-invasive tests disagree, the diagnosis would change treatment, or enrollment in a clinical trial requires histological confirmation.

What Score Changes Mean for Treatment

If you have been diagnosed with steatohepatitis, one of the clearest interventions is weight loss. Research has shown that losing at least 5% of body weight produces measurable drops in the NAS, and that achieving 10% or more weight loss leads to NAS improvement in virtually all patients, resolution of steatohepatitis in about 90%, and regression of fibrosis in about 45%. The catch is that only about 10% of study participants managed to lose that much weight.²¹

In clinical drug trials, the NAS plays a formal role in defining treatment success. Regulatory guidance has established two primary endpoints for drugs aimed at fatty liver disease: resolution of steatohepatitis without worsening fibrosis, or at least a two-point improvement in the NAS with improvement in more than one category and no worsening of fibrosis.²² A joint workshop by major liver disease associations reinforced that either resolution of steatohepatitis or improvement in fibrosis, assessed on paired biopsies, remains the accepted standard for drug approval.²³ So while the NAS may not predict your long-term survival as powerfully as fibrosis stage does, it remains the yardstick for judging whether a treatment is working in the short term.

Artificial Intelligence and the Future of Liver Scoring

One of the most frustrating aspects of biopsy-based scoring is that different pathologists can look at the same tissue and reach different conclusions. The subjectivity built into grading ballooning or counting inflammatory foci introduces variability that muddies both clinical care and research. Machine learning models trained on thousands of digitized biopsy slides are beginning to address this. In an analysis of a semaglutide trial, an AI pathology model was compared against conventional pathologist reads for scoring the histological features of steatohepatitis.²⁴

Validation testing of these algorithms has shown high agreement rates for the key histological features, with repeatability scores ranging from 0.93 to 0.96 across different scanners at different locations, outperforming the repeatability of human pathologist reads.²⁵ If these tools become standard, the NAS derived from a biopsy could become far more consistent and trustworthy than it is today, which would benefit both individual patients wondering whether their disease has improved and clinical trials trying to prove whether a new drug works.

1
PubMed Central. Nonalcoholic fatty liver disease (NAFLD) activity score and the histopathologic diagnosis in NAFLD: distinct clinicopathologic meanings
2
Elsevier / ScienceDirect (Journal of Hepatology). A multisociety Delphi consensus statement on new fatty liver disease nomenclature
3
MDPI. Assessing Liver Fibrosis Using the FIB4 Index in the Community Setting
4
JAMA Network Open. Diagnostic Performance of the Fibrosis-4 Index and Nonalcoholic Fatty Liver Disease Fibrosis Score in Lean Adults With Nonalcoholic Fatty Liver Disease
5
Europe PMC. NAFLD fibrosis score: a prognostic predictor for mortality and liver complications among NAFLD patients
6
Cureus. Real-World Feasibility of Fibrosis-4-Based Fibrosis Risk Stratification in Patients With Type 2 Diabetes Mellitus and Metabolic Dysfunction-Associated Steatotic Liver Disease
7
Wiley Online Library. Enhanced Liver Fibrosis (ELF) Test for Identifying Significant Fibrosis, Advanced Fibrosis, and Cirrhosis in Patients With Metabolic Dysfunction-Associated Steatotic Liver Disease
8
PubMed Central. Enhanced Liver Fibrosis Score for Diagnosing Liver Fibrosis in Chronic Hepatitis
9
PubMed Central. Association of dietary fiber sources with liver fibrosis assessed by transient elastography in U.S. adults: a cross-sectional study from NHANES 2017 to 2020
10
PubMed Central. Accuracy of controlled attenuation parameter (CAP) and liver stiffness measurement (LSM) for assessing steatosis and fibrosis in non-alcoholic fatty liver disease: A systematic review and meta-analysis
11
Frontiers in Medicine. Diagnostic performance of controlled attenuation parameter for grading hepatic steatosis in MASLD: an MRI-PDFF-referenced study in a Chinese cohort
12
CrossRef. Performance of Proton Density Fat Fraction, MRE-based Liver Stiffness, and T1 in Identifying NASH Patients at High Risk of Disease Progression
13
The Lancet. FibroScan-AST (FAST) score for the non-invasive identification of patients with non-alcoholic steatohepatitis with significant activity and fibrosis: a prospective derivation and global validation study
14
New England Journal of Medicine. Prospective Study of Outcomes in Adults with Nonalcoholic Fatty Liver Disease
15
Wiley Online Library. Increased risk of mortality by fibrosis stage in non‐alcoholic fatty liver disease: Systematic Review and Meta‐analysis
16
Elsevier / Clinical Gastroenterology and Hepatology. Mortality Outcomes by Fibrosis Stage in Nonalcoholic Fatty Liver Disease: A Systematic Review and Meta-analysis
17
Elsevier / Journal of Hepatology. Fibrosis stage but not NASH predicts mortality and time to development of severe liver disease in biopsy-proven NAFLD
18
PubMed Central. Sampling variability of liver biopsy in nonalcoholic fatty liver disease
19
Dove Medical Press. Identifying High-Risk NASH Patients: What We Know so Far
20
Europe PMC. Nonalcoholic Fatty Liver Disease Screening in Type 2 Diabetes Mellitus Patients in the Primary Care Setting
21
PubMed Central. Nonalcoholic Fatty Liver Disease and Obesity Treatment
22
PubMed Central. Endpoints and Clinical Trial Design for Nonalcoholic Steatohepatitis
23
Hepatology. Report on the AASLD/EASL Joint Workshop on Clinical Trial Endpoints in NAFLD
24
Wolters Kluwer Health. Artificial intelligence scoring of liver biopsies in a phase II trial of semaglutide in nonalcoholic steatohepatitis
25
PubMed Central. Artificial intelligence-assisted digital pathology for non-alcoholic steatohepatitis: current status and future directions