ARE YOU SAFE?
SAFE is a score that estimates the likelihood of clinically significant liver fibrosis using routinely available health information and laboratory tests.
SAFE was developed for metabolic dysfunction-associated steatotic liver disease (MASLD) to identify clinically significant fibrosis (F2 or higher).
It uses seven model variables: age, body mass index (BMI), diabetes, and the following laboratory data: AST, ALT, platelets, and globulin (total protein minus albumin). These latter laboratory data are parts of routinely used blood tests including complete blood count (CBC) and hepatic function panel or comprehensive metabolic panel (CMP).
For patients
Understand what MASLD and liver fibrosis are, why fibrosis matters, what patients can do, and where SAFE fits in identifying people who may need further assessment.
For practitioners
Learn what SAFE measures, how to use and interpret it, how well it performs compared with FIB-4, and the key warnings, limitations, and clinical evidence.
For public health
Explore first-line fibrosis screening, testing capacity, and how SAFE and FIB-4 compare when the same proportion of screening candidates undergo transient elastography.
Liver fibrosis points to liver disease severe enough to increase the future risk of poor outcomes, such as liver failure.
The good news is that it also identifies patients who may benefit from approved medical therapies.
Fibrosis is scarring caused by ongoing liver injury.
F1 is minimal fibrosis. F2–4 represents significant fibrosis, where F4 is cirrhosis.
Sometimes F2-F3 is referred to as moderate fibrosis and patients with F2-F3 are primary candidates for approved medical therapies including semaglutide and resmetirom.
Understanding your liver health
SAFE is designed to help identify people who may have clinically significant liver fibrosis (scarring of the liver from persistent liver injury) from metabolic dysfunction-associated steatotic liver disease, which used to be called non-alcoholic fatty liver disease (NAFLD).
What is MASLD?
Why does fat accumulate in the liver?
Insulin resistance is central to MASLD. It is closely linked with obesity, type 2 diabetes, abnormal lipids and other cardiometabolic conditions, which causes the liver to produce excess fat.
Obesity
Excess adipose tissue, especially in the truncal region, and metabolic dysfunction increase the likelihood of liver fat and progressive disease.
Insulin resistance & diabetes
Abnormal glucose metabolism is a central part of MASLD risk and is associated with greater risk of fibrosis.
Cardiometabolic health
High blood pressure and abnormal blood lipids are part of the same metabolic abnormalities that also elevates cardiovascular risk and contributes importantly to overall health.
Fat is not the whole story.
MASLD is very common.
Roughly a third of the U.S. and global population has MASLD. Good news is that most people with MASLD will not develop advanced liver disease. That is precisely why risk stratification matters: we need to identify the smaller group in whom liver injury is progressing.
Symptoms and routine liver tests do not reliably tell you how much fibrosis is present. That is why fibrosis risk assessment is important before symptoms of advanced disease appears.
The goal is to find clinically significant fibrosis before liver disease becomes advanced.
SAFE is a first-pass risk-stratification tool designed for the F2+ question. It does not diagnose fibrosis by itself; it helps identify who may benefit from further assessment.
Liver injury can progress from fat accumulation to inflammation, fibrosis, and cirrhosis.
What can patients do?
Treatment starts with improving the metabolic drivers of liver disease and overall health. The exact plan should be individualized with a healthcare professional. Here are some general recommendations.
Assess liver fibrosis
People at increased metabolic risk may benefit from noninvasive fibrosis assessment to determine whether additional testing is appropriate.
Treat MASLD with fibrosis
Fibrosis stage is the main parameter clinicians use to determine whether a patient with MASLD may be eligible for liver-directed pharmacotherapy. Treatment decisions should incorporate fibrosis stage along with the overall clinical context.
Improve metabolic health
Weight management remains a key component in MASLD treatment. Depending on the clinical status, 5-10% weight loss is recommended. Healthy eating patterns and regular physical activity, including aerobic and muscle-toning exercises, as well as proper management of diabetes, blood pressure and abnormal lipids are central elements in MASLD care.
Protect the liver
Review alcohol use, medications and other causes of liver injury with a clinician.
Using SAFE in clinical practice
SAFE is a blood-based clinical score developed to estimate the likelihood of clinically significant liver fibrosis (F2–F4) in people with MASLD using routinely available clinical and laboratory data.
Find significant fibrosis early enough to act.
SAFE was developed to distinguish F0–F1 from F2–F4 fibrosis. F2 is the beginning of clinically significant fibrosis and is an important threshold for treatment eligibility in appropriately selected patients.
Seven model variables
Age, BMI, diabetes, AST, ALT, platelet count, and globulin.
Eight calculator inputs
Globulin is calculated automatically from total protein minus albumin, so the calculator asks for both values separately.
Development and biopsy validation: Sripongpun et al., Hepatology 2023. PubMed ↗
Use SAFE as a first-line risk assessment—not as a diagnosis.
The website calculator reports the SAFE score together with an estimated probability of elevated liver stiffness. The probability estimates are calibrated to transient elastography in NHANES and are intended to help interpret the score continuously rather than reducing it to a single category.
SAFE <0 has been proposed to designate a low probability of liver fibrosis and a very low probability of cirrhosis. Increasing SAFE scores are associated with progressively higher probabilities of liver fibrosis and elevated liver stiffness. SAFE should therefore be interpreted as a continuous risk estimate rather than as a diagnosis.
Threshold selection depends on the clinical setting and available testing capacity: lower thresholds detect more disease but send more people for second-level testing; higher thresholds reduce testing but miss more disease.
SAFE has been validated against fibrosis and compared with other blood-based scores.
The original biopsy-based development and validation studies reported AUROCs of at least 0.80 for distinguishing F0–F1 from ≥F2. Subsequent U.S. population and Asian studies independently validated SAFE, and multiple studies have reported better discrimination for significant fibrosis than FIB-4.
| SAFE | FIB-4 | |
|---|---|---|
| Primary fibrosis target | Clinically significant fibrosis beginning at F2 | Historically used primarily for advanced fibrosis risk |
| Variables | Age, BMI, diabetes, AST, ALT, platelets, globulin | Age, AST, ALT, platelets |
In the U.S. external validation study, SAFE showed greater discrimination for clinically significant fibrosis than FIB-4.
Interpret SAFE in context.
Risk is not diagnosis
SAFE estimates probability. Confirmation of fibrosis requires an appropriate second-level assessment such as transient elastography, another validated fibrosis test, or histology when clinically indicated.
Population matters
Predictive values depend on disease prevalence. A probability calibrated in a population sample may differ from the probability in a hepatology referral population.
No threshold is perfect
Every cutoff trades sensitivity against testing burden. That is why the SAFE calculator presents risk continuously as estimated probabilities rather than relying only on categorical thresholds.
Age
SAFE has been less robustly validated at the extremes of age, and interpretation warrants additional caution in very young and very old adults.
SAFE has also shown prognostic associations with mortality, liver-related events, and HCC, but prognostic performance answers a different question from identifying F2 fibrosis. Hong Kong study ↗ · HCC study ↗
Explore SAFE in more depth.
SAFE calculator
Enter routine clinical and laboratory data to calculate SAFE and view continuous probability estimates.
Screening strategy
Explore how SAFE and FIB-4 compare when the same proportion of screening candidates undergo transient elastography.
Clinical evidence
Primary publications supporting SAFE development, validation, comparative performance, and prognosis.
Sripongpun et al. Development and biopsy validation. Hepatology. PubMed ↗
van Kleef et al. U.S. population validation. Hepatology Communications. PubMed ↗
Diagnostic and prognostic performance in Hong Kong NAFLD cohorts. PubMed ↗
Machine-learning benchmark for significant fibrosis. Hepatology. PubMed ↗
Population screening comparison with FIB-4. Annals of Hepatology. PubMed ↗
SAFE and hepatocellular carcinoma risk. Clinical and Molecular Hepatology. PubMed ↗
External validation of SAFE in 6,229 Asian patients with MASLD. Yonsei Medical Journal. PubMed ↗
Calculate SAFE.
Eight familiar inputs supply the seven variables in SAFE. The score is then translated into continuous, NHANES-calibrated probabilities of elevated liver stiffness, with bootstrap 95% confidence intervals.
Privacy: Your calculator information stays in your browser and is not transmitted or stored by LiverSAFE.
Screening is a resource-allocation problem.
Adults with type 2 diabetes or obesity are important candidates for MASLD fibrosis screening. Using BMI ≥30 kg/m²—or ≥25 kg/m² for Asian adults—this represents approximately 47.5% of the U.S. adult population in NHANES 2017–March 2020.
If SAFE or FIB-4 is used as a first-line test followed by transient elastography, which strategy does a better job?
At a given transient-elastography capacity, see how the two scores compare when the same percentage of people are selected for testing.
SAFE
FIB-4
NHANES 2017–March 2020 pre-pandemic screening candidates, N=3,631: type 2 diabetes or obesity (BMI ≥30 kg/m²; BMI ≥25 kg/m² for non-Hispanic Asian adults). Screening candidates represent approximately 47.5% of U.S. adults using NHANES examination weights. The simulator uses the unweighted complete-case analytic sample.
Resources for patients and clinicians.
For patients
MASLD, fibrosis, what a SAFE result means, and questions to discuss with a healthcare professional.
For clinicians
Screening concepts, FIB-4 limitations, SAFE interpretation, confirmatory assessment and clinical caveats.
Scientific resources
Publications, methodology, disclosures, development history and future evidence updates.