Liver fibrosis is an excessive accumulation of extracellular scar tissue in the liver, usually resulting from persistent or recurrent injury. It is a form of fibrosis in which wound-healing responses become sustained, altering the composition and organization of the tissue surrounding liver cells and blood vessels. Common causes include chronic viral hepatitis, alcohol-associated liver disease, metabolic dysfunction-associated steatotic liver disease, autoimmune liver disease, and disorders that obstruct or damage bile ducts.
Fibrosis is clinically important because its extent is closely associated with the risk of future liver-related illness and death. Advanced fibrosis can progress to cirrhosis, characterized by extensive scarring, regenerative nodules, and distortion of the liver’s vascular architecture. However, fibrosis and cirrhosis are not interchangeable terms: early fibrosis may leave much of the liver’s structure and function intact, and even substantial fibrosis can sometimes regress when the underlying injury is controlled.
The condition is assessed through tissue examination, blood-based tests, and imaging techniques that estimate liver stiffness. These methods measure different aspects of disease and are not perfectly interchangeable. Modern understanding treats fibrosis as a dynamic biological process involving liver cells, immune responses, extracellular matrix turnover, and mechanical changes in tissue, rather than as an inert, permanently fixed scar.
Tissue architecture and the meaning of fibrosis
The liver’s microscopic organization explains why fibrosis affects both metabolism and circulation. Blood from the portal vein and hepatic artery enters small vascular channels called sinusoids, passes alongside plates of hepatocytes, and drains toward central veins. Hepatocytes carry out much of the liver’s biochemical work, including processing nutrients, synthesizing plasma proteins, and modifying substances for elimination. Between the sinusoidal endothelial lining and hepatocytes lies the space of Disse, a narrow compartment through which substances are exchanged.
In a healthy liver, the extracellular matrix provides a relatively delicate supporting framework. It contains collagens, glycoproteins, and proteoglycans arranged differently across the liver’s compartments. Fibrosis involves both an increase in matrix quantity and a change in matrix composition. Fibrillar collagens, particularly types I and III, accumulate in locations where they are normally sparse. Fibronectin, laminin, and other matrix components also change, affecting how cells attach, migrate, and respond to signals.
A characteristic accompanying alteration is sinusoidal capillarization. Normal sinusoidal endothelial cells have small openings, called fenestrae, and lack the continuous basement membrane typical of many other capillaries. With chronic injury, fenestrae are lost and basement-membrane-like material accumulates. These changes impede exchange between blood and hepatocytes. Thus, fibrosis can alter liver physiology before the entire organ becomes nodular or conventional measures of liver function become abnormal.
The distribution of scar tissue depends partly on the location of the original injury. Chronic viral hepatitis commonly produces fibrosis around portal tracts, which contain branches of the portal vein, hepatic artery, and bile duct. Alcohol-associated and metabolic steatohepatitis often begin with perisinusoidal and pericellular fibrosis around central veins. Biliary disorders primarily produce portal and periportal changes. As disease advances, fibrous bands can connect portal tracts to one another or to central veins, producing bridging fibrosis.
Scar quantity alone does not fully describe disease severity. The thickness and connectivity of fibrous bands, the presence of regenerative nodules, vascular remodeling, and ongoing inflammatory activity all matter. Congenital hepatic fibrosis also illustrates the distinction between fibrosis and cirrhosis: this developmental disorder causes portal fibrosis and abnormal bile duct structures, often with portal hypertension, but does not necessarily produce the same diffuse hepatocellular injury and nodular architecture as acquired cirrhosis. Liver fibrosis is therefore a tissue response shared by multiple diseases, not a single disease with one uniform microscopic pattern.
Causes and modifying factors
Persistent liver injury can arise through infectious, metabolic, toxic, immune-mediated, biliary, vascular, or inherited mechanisms. Worldwide, hepatitis B and hepatitis C have been major causes. In hepatitis B, much of the injury reflects immune responses against infected hepatocytes rather than direct destruction by the virus alone. Chronic hepatitis C produces sustained inflammatory injury through interactions among viral infection, hepatocyte stress, and host immunity. The pace of scarring varies widely, and infection duration does not translate into a fixed fibrosis stage.
Metabolic dysfunction-associated steatotic liver disease, or MASLD, is associated with hepatic fat accumulation and cardiometabolic risk factors. It falls within the broader subject of fatty liver disease. The term metabolic dysfunction-associated steatohepatitis, or MASH, describes a form with hepatocellular injury and inflammation. The MASLD and MASH terminology was introduced in a 2023 multisociety nomenclature process, replacing NAFLD and NASH in much clinical and research usage. Obesity and type 2 diabetes are important associated conditions, although advanced metabolic liver disease also occurs in people without obesity. Fibrosis can persist when the amount of visible liver fat decreases, particularly in advanced disease.
Alcohol-associated liver disease results from interacting effects of ethanol metabolism, oxidative stress, altered gut permeability, inflammatory signaling, and nutritional or metabolic factors. Acetaldehyde, a product of ethanol metabolism, contributes to cellular injury and fibrogenic signaling. Alcohol exposure and metabolic dysfunction may coexist, and their contributions cannot always be separated into independent categories. The amount and duration of exposure are important, but individual susceptibility also varies substantially.
Immune-mediated causes include autoimmune hepatitis, primary biliary cholangitis, and primary sclerosing cholangitis. Autoimmune hepatitis primarily involves inflammatory injury to hepatocytes. The latter two disorders prominently affect bile ducts and can cause retention of bile constituents, ductular reactions, and progressive portal scarring. Inherited conditions include hereditary hemochromatosis, Wilson disease, alpha-1 antitrypsin deficiency, and some disorders of intermediary metabolism. Certain medications and environmental exposures can also produce chronic injury and fibrosis through distinct mechanisms.
Vascular and infectious causes create additional patterns. Chronic hepatic congestion from heart failure initially affects regions around central veins, while hepatic venous outflow obstruction can produce congestion, hepatocyte loss, and scarring. Schistosomiasis characteristically causes periportal fibrosis and may produce severe portal hypertension despite relatively preserved hepatocyte function. These examples show why the presence of fibrosis alone does not establish its cause.
Progression is modified by age, metabolic disease, ongoing alcohol exposure, coinfections, and inherited susceptibility. Variants in genes such as PNPLA3 and TM6SF2 influence risk in steatotic liver disease, but do not determine an individual outcome. Multiple causes often coexist, making etiologic assessment more complex than assigning every case to a single category.
Cellular mechanisms of scar formation
The central cellular event in many forms of liver fibrosis is activation of the hepatic stellate cell. In its resting state, this cell lies in the space of Disse and stores vitamin A in lipid droplets. Following injury, it undergoes extensive changes in gene expression, loses much of its lipid-storage phenotype, proliferates, and acquires contractile and matrix-producing properties resembling those of a myofibroblast. Portal fibroblasts and other mesenchymal populations also contribute, particularly in biliary injury.
Activation is driven by signals from damaged hepatocytes, resident macrophages, recruited immune cells, endothelial cells, and the extracellular matrix itself. Dying hepatocytes release molecular signals that stimulate inflammation. Kupffer cells, the liver’s resident macrophages, and recruited monocyte-derived macrophages produce mediators that influence both injury and repair. Transforming growth factor beta is a major stimulus for collagen production, while platelet-derived growth factor promotes the proliferation and migration of activated mesenchymal cells.
The matrix accumulates when its production exceeds its breakdown. Matrix metalloproteinases can degrade extracellular matrix components, but their activity is regulated by tissue inhibitors of metalloproteinases. During sustained injury, this balance often favors matrix preservation. Collagen also undergoes chemical cross-linking, including reactions mediated by lysyl oxidase family enzymes, which increases tissue stiffness and can make established scar less susceptible to enzymatic degradation.
Mechanical changes then reinforce biological signaling. Cells detect a stiffened matrix through integrins, cytoskeletal structures, and mechanosensitive signaling pathways. These signals can maintain myofibroblast activation and alter hepatocyte and endothelial behavior. Fibrosis consequently has a self-reinforcing component: injury promotes matrix deposition, and the changed matrix helps sustain the fibrogenic state. This does not mean progression becomes inevitable, but it helps explain why removing an initiating cause may not immediately restore normal architecture.
The immune system has both profibrotic and restorative functions. Particular macrophage states promote inflammation and stellate-cell activation, whereas others contribute to matrix degradation and resolution. Natural killer cells can kill some activated stellate cells, while different T-cell populations influence inflammation, cell survival, and repair. These functions vary with disease stage and tissue context; a simple division into “harmful” and “protective” immune cells is inadequate.
Additional influences include oxidative stress, altered lipid metabolism, cellular senescence, and communication between the intestine and liver. Microbial products reaching the liver through portal blood can activate innate immune pathways, especially when intestinal barrier function is disturbed. These overlapping mechanisms help explain why a treatment effective against one signaling pathway may have limited effects in established human disease. Fibrosis is maintained by a network of interacting cell populations and signals rather than by one universally dominant molecular switch.
Histological assessment and staging
Examination of a liver biopsy can demonstrate fibrosis directly while also identifying features of the underlying disease. Tissue sections show where scar is located, whether fibrous bridges are present, and whether regenerative nodules have formed. Routine hematoxylin and eosin staining provides information about cellular injury and tissue organization, while connective-tissue stains such as Masson trichrome or picrosirius red make collagen more conspicuous. Reticulin staining helps assess the supporting framework and changes in hepatocyte plate architecture.
Fibrosis stage is distinct from inflammatory grade. Stage describes the extent and architectural distribution of scarring; grade describes inflammatory activity and associated tissue injury. A person may have substantial established fibrosis with little current inflammation, or active hepatitis with relatively limited fibrosis. Confusing these dimensions can obscure the difference between accumulated structural damage and the intensity of the injury occurring at the time of sampling.
Several staging systems are used. METAVIR, developed for chronic hepatitis C, assigns fibrosis stages from F0 to F4: F0 indicates no fibrosis, F1 portal fibrosis without septa, F2 portal fibrosis with a limited number of septa, F3 numerous septa without cirrhosis, and F4 cirrhosis. The Ishak system uses stages from 0 to 6, allowing more subdivisions of bridging fibrosis and cirrhosis. These scales are related but not interchangeable, and numerical changes do not represent equal increments of collagen content.
Steatohepatitis requires a system attentive to its characteristic distribution. The NASH Clinical Research Network system separates early perisinusoidal or portal patterns within stage 1, identifies combined perisinusoidal and portal/periportal fibrosis as stage 2, bridging fibrosis as stage 3, and cirrhosis as stage 4. Terminology such as “significant fibrosis,” “advanced fibrosis,” and “cirrhosis” often corresponds to F2 or greater, F3 or greater, and F4, respectively, in four-stage systems. These definitions depend on the scale and study context.
Biopsy has important limitations. Fibrosis is unevenly distributed, and a small sample represents only a minute portion of the liver. A specimen that misses a fibrous bridge or a nodule may underestimate stage, while an unusually scarred fragment may overestimate the wider burden. Specimen length, fragmentation, and the number of portal tracts affect interpretability. Observer variation also occurs, especially between adjacent stages.
Because biopsy is invasive and carries risks including pain and bleeding, its role has changed as noninvasive tests have improved. It remains especially informative when the cause of disease is uncertain, several conditions may coexist, or microscopic inflammatory features influence treatment classification. Quantitative collagen measurements and digital pathology can supplement conventional staging, but a collagen percentage alone cannot fully capture vascular distortion, scar location, or the biological activity of the disease.
Natural history and clinical consequences
Liver fibrosis usually develops without a specific symptom attributable to the scar itself. Early disease may be discovered during investigation of abnormal liver enzymes, incidental imaging findings, or assessment of a known chronic liver disorder. Normal aminotransferase concentrations do not exclude substantial fibrosis. Conversely, markedly elevated enzymes may reflect active injury without advanced scarring. Symptoms such as fatigue are nonspecific and do not reliably indicate fibrosis stage.
Progression is heterogeneous. Some people remain at an early stage for many years, whereas others develop advanced disease over a shorter interval. The process is not necessarily linear: periods of active injury can alternate with stabilization or regression. Estimates expressed as “fibrosis stages per year” are population-level summaries and can be misleading when treated as a personal timetable. Age, cause of disease, metabolic factors, and the persistence or removal of injury all influence the trajectory.
As scar and vascular remodeling increase, resistance to portal blood flow rises. Portal hypertension has both structural and dynamic components: fibrous septa and nodules obstruct flow, while activated contractile cells and altered vascular signaling further increase resistance. In sinusoidal liver disease, a hepatic venous pressure gradient of at least 10 mmHg defines clinically significant portal hypertension and is associated with an increased risk of complications. This measurement is less informative in predominantly presinusoidal disorders, in which portal pressure may be high without an equivalently elevated gradient.
Cirrhosis can remain compensated, meaning that major clinical complications have not yet occurred. Decompensation is marked by events such as ascites, variceal bleeding, or hepatic encephalopathy. Ascites involves fluid accumulation in the abdominal cavity through interactions among portal hypertension, altered circulatory regulation, and renal sodium retention. Varices are collateral veins that develop as portal blood is diverted around the liver. Encephalopathy reflects brain dysfunction associated with liver insufficiency and portosystemic shunting, rather than a direct mechanical effect of scar tissue.
Advanced fibrosis also changes the risk of liver cancer, especially hepatocellular carcinoma. Cirrhosis is a major risk state, although cancer can occur without cirrhosis in certain diseases, notably chronic hepatitis B. Fibrosis is not itself a malignant lesion. Its association with cancer reflects the wider environment of chronic injury, regeneration, inflammation, genetic alterations, and tissue remodeling.
Across several chronic liver diseases, fibrosis stage is a strong predictor of liver-related outcomes. Interpretation nevertheless requires attention to competing risks. In metabolic liver disease, cardiovascular disease and nonhepatic cancers contribute substantially to illness and death. Measures such as the Child–Pugh classification and MELD score describe aspects of advanced liver disease severity and prognosis; they are not direct fibrosis-staging systems. Scar burden, functional reserve, portal pressure, and overall health are related but separate dimensions.
Blood-based tests and sequential assessment
Blood-based fibrosis tests use either indirect indicators of chronic liver disease or components associated with matrix turnover. Indirect markers include platelet count, aminotransferases, and patient age. Direct-marker panels measure substances such as hyaluronic acid or proteins involved in collagen metabolism. A biomarker can correlate with fibrosis without being specific to scar tissue, so each test must be interpreted in relation to the population and clinical setting in which it was validated.
The fibrosis-4 index, usually called FIB-4, combines age, aspartate aminotransferase, alanine aminotransferase, and platelet count:
The conventional calculation uses age in years, AST and ALT in units per liter, and platelets in . Platelet reduction can accompany portal hypertension and splenic sequestration, while aminotransferases provide information about liver injury. However, neither mechanism is unique to fibrosis. Platelet disorders, acute hepatitis, and changes in enzyme activity can therefore alter the result independently of scar burden.
FIB-4 is particularly useful as an initial risk-stratification tool rather than as an exact histological stage measurement. In many adult MASLD pathways, a value below 1.3 identifies a group with a relatively low probability of advanced fibrosis, while values above 2.67 indicate a higher probability. The interval between these values is indeterminate. Age-adjusted interpretation is important: a lower decision boundary of 2.0 is commonly used in people older than 65 years, and performance is limited in adults younger than 35. These thresholds are context-dependent, not universal diagnostic boundaries.
The AST-to-platelet ratio index, or APRI, is another widely used score, particularly in viral hepatitis. It combines AST relative to the laboratory’s upper limit of normal with platelet count. Proprietary panels combine multiple biochemical measurements. The Enhanced Liver Fibrosis test measures hyaluronic acid, the amino-terminal propeptide of type III procollagen, and tissue inhibitor of metalloproteinases 1. Such tests can provide information about prognosis as well as the likelihood of advanced fibrosis, but may also be affected by biological processes outside the liver.
Test performance depends on disease prevalence and the purpose of testing. In a low-prevalence population, a low-risk result may be useful for excluding advanced disease, while a positive result may require confirmation. Sensitivity and specificity do not by themselves determine the probability that a particular result represents true advanced fibrosis. Predictive values change with the pretest likelihood of disease.
Sequential approaches combine a relatively inexpensive first-line blood score with a second test, commonly elastography or a specialized serum panel. This reduces the number of people requiring invasive assessment while identifying groups at higher risk. Discordant results remain important: a low blood score and high stiffness measurement may reflect a confounder, measurement limitations, or disease not well represented by one test. Noninvasive scores classify probabilities and risks; they do not create a universally precise substitute for microscopic staging.
Imaging and liver stiffness measurement
Conventional ultrasonography, computed tomography, and magnetic resonance imaging can show features associated with advanced chronic liver disease, including a nodular liver surface, altered proportions of liver segments, enlarged spleen, collateral vessels, and ascites. These findings are less sensitive for early fibrosis. A liver that appears unremarkable on routine imaging can still contain clinically important scar tissue, while changes in contour or texture do not establish the underlying cause.
Elastography assesses mechanical properties of tissue. In vibration-controlled transient elastography, an external mechanical impulse generates a shear wave, and ultrasound tracks its propagation through the liver. Shear waves generally travel faster in stiffer tissue. The result is usually expressed in kilopascals as an estimated stiffness. Point shear-wave and two-dimensional shear-wave elastography use acoustic impulses generated by the ultrasound system, allowing stiffness assessment alongside conventional imaging.
Stiffness is not a direct measurement of collagen. Active inflammation, extrahepatic biliary obstruction, hepatic congestion, and recent food intake can increase stiffness independently of fibrosis. Technical factors, including body habitus, probe selection, measurement location, and acquisition quality, also influence results. Conventional vibration-controlled transient elastography is generally limited by ascites, whereas some other ultrasound-based techniques can be used in its presence. Appropriate interpretation therefore requires both technical quality assessment and knowledge of the clinical context.
Thresholds differ among devices and disease populations. In selected chronic liver disease settings, vibration-controlled transient elastography values below approximately 8 kPa are often used to help exclude advanced fibrosis. The Baveno VII consensus uses a related but distinct framework for compensated advanced chronic liver disease: values below 10 kPa can rule it out when other clinical or imaging signs are absent, 10–15 kPa are suggestive, and values above 15 kPa are highly suggestive. These categories concern clinical risk and advanced disease, not an exact one-to-one conversion into biopsy stages.
Magnetic resonance elastography generates mechanical waves and uses MRI sequences to map their propagation over larger portions of the liver. It generally has high diagnostic accuracy for advanced fibrosis and can be useful when ultrasound-based measurements are technically unsuccessful or inconclusive. Limitations include cost, availability, and technical difficulties in some settings, including substantial iron overload with certain sequences. Its kilopascal values are not interchangeable with those from transient elastography because the methods estimate tissue properties differently.
Repeated imaging can document changes over time, but a fall in stiffness after treatment does not necessarily prove equivalent removal of scar. For example, antiviral therapy may rapidly reduce inflammation, producing a stiffness reduction before substantial collagen remodeling occurs. Imaging is therefore most informative when integrated with laboratory findings, disease activity, and prior measurements obtained under comparable conditions. Structural imaging, stiffness measurement, and histology provide overlapping but nonidentical information.
Treatment through control of the underlying disease
Treatment of liver fibrosis has historically depended primarily on controlling the cause of ongoing injury. This differs from treating a symptom of fibrosis or directly dissolving collagen. Removal or suppression of the initiating injury can reduce inflammatory signaling, decrease stellate-cell activation, and permit matrix remodeling. The degree of subsequent recovery depends on baseline architecture, duration of disease, associated conditions, and whether vascular or regenerative changes have become established.
In chronic hepatitis C, direct-acting antiviral drugs can achieve sustained virologic response in most treated patients. Viral eradication is associated with reduced liver-related complications and with regression of fibrosis in some patients. In chronic hepatitis B, nucleos(t)ide analogues such as entecavir and tenofovir suppress viral replication and can produce histological improvement, including regression of cirrhosis in some long-term studies. Suppression is not equivalent to eradication of hepatitis B, and the clinical implications differ from cure of hepatitis C.
Control of alcohol-associated injury centers on cessation of the injurious exposure and treatment of alcohol use disorder within broader medical care. In metabolic liver disease, sustained weight reduction and improvement in metabolic dysfunction have been associated with improvement in steatohepatitis and, in some patients, fibrosis. Effects on liver fat, inflammatory activity, and fibrosis occur on different timescales. A drug that improves glycemic control or reduces hepatic fat does not automatically have a demonstrated antifibrotic effect.
A notable regulatory development occurred in March 2024, when the United States Food and Drug Administration granted accelerated approval to resmetirom for adults with noncirrhotic NASH, now commonly termed MASH, with moderate-to-advanced fibrosis corresponding to F2–F3, alongside diet and exercise. Resmetirom is a thyroid hormone receptor beta agonist. The approval was based on histological endpoints, with confirmatory evidence required to establish clinical benefit. This was an indication for a defined disease and fibrosis range, not a general treatment for fibrosis from every cause or for decompensated cirrhosis.
Other etiologies require distinct approaches. Immunosuppressive treatment can control autoimmune hepatitis; ursodeoxycholic acid is established treatment for primary biliary cholangitis; iron removal treats iron overload in hereditary hemochromatosis; and copper-directed treatment addresses Wilson disease. Relief of a reversible biliary obstruction can reduce cholestatic injury. These examples illustrate why identifying the cause remains necessary even when fibrosis has already been demonstrated.
Management of established advanced disease also addresses portal hypertension, ascites, encephalopathy, infection risk, and liver cancer risk. Such care is not synonymous with reversing fibrosis. Liver transplantation replaces the diseased organ in selected cases of end-stage liver disease, but is not a treatment for uncomplicated early fibrosis. Clinical outcomes depend on both etiologic control and management of the structural and functional consequences already present.
Regression, antifibrotic research, and unresolved measurement problems
Fibrosis regression is biologically possible because extracellular matrix is continually produced, modified, and degraded. When injury diminishes, some activated stellate cells undergo programmed cell death, some become senescent, and others enter a less active state. Macrophage populations can shift toward functions that support matrix breakdown and tissue repair. Collagen-degrading enzymes then act within a changed balance of matrix production and inhibition.
Regression is not equivalent to a complete return to normal. Thick, heavily cross-linked septa may persist, and vascular shunts or distorted microcirculation can remain after substantial collagen loss. Histological features of regression include thinning or fragmentation of fibrous septa and changes in the relationship between scar and regenerating tissue. Even when a biopsy no longer meets all criteria for cirrhosis, residual architectural and vascular abnormalities may retain clinical significance.
Human evidence for reversibility became particularly influential in the late twentieth and early twenty-first centuries. Studies following successful hepatitis C treatment and prolonged hepatitis B suppression demonstrated that established fibrosis could decrease. Improvements after treatment of some metabolic, autoimmune, and iron-overload disorders provided additional evidence. Interpretation of paired biopsies nevertheless remains affected by sampling variability: an apparent stage change in one person may partly reflect differences in the tissue sampled.
Direct antifibrotic research has targeted several stages of scar formation. Investigational strategies include reducing stellate-cell activation, modifying macrophage recruitment, inhibiting collagen cross-linking, altering metabolic stress, and promoting matrix degradation. Broad inhibition of transforming growth factor beta is difficult because this pathway also participates in normal immune regulation, tissue maintenance, and tumor suppression. Simtuzumab, an antibody targeting lysyl oxidase-like 2, did not demonstrate the expected clinical benefit in major trials in steatohepatitis and primary sclerosing cholangitis, illustrating the gap between plausible mechanisms and effective treatment.
Measurement is a major obstacle in clinical trials. Clinical events such as decompensation, transplantation, and liver-related death may take years to accumulate, especially in earlier disease. Trials therefore often use histological improvement or noninvasive markers as intermediate outcomes. A commonly used biopsy endpoint is improvement in fibrosis by at least one stage without worsening steatohepatitis. Such an endpoint is a surrogate endpoint: its usefulness depends on how reliably treatment-induced changes predict meaningful clinical benefits.
Emerging methods include quantitative image analysis, spatial mapping of cell populations, collagen-turnover markers, and combinations of blood and imaging measurements. Their validation requires comparisons across etiologies, stages, laboratories, and demographic groups. A central unresolved problem is distinguishing reduced disease activity from true architectural repair. Another is determining whether a measured reduction in scar produces a proportional reduction in portal hypertension, decompensation, or cancer risk. These are related outcomes, but existing evidence does not support treating them as interchangeable.