Cirrhosis is an important cause of morbidity and mortality. In the United States, cirrhosis accounts for more than 33,000 deaths per year and is the 12th leading cause of death. It is reported on hospital discharge in at least 1% of adult patients and is found in 4% to 12% of patients at autopsy in developed countries. Cirrhosis is the morphologic result of many different types of chronic insult to the liver. It may develop rapidly during a period of months, but most often it is a product of many years of chronic injury. The causes include chronic viral or autoimmune hepatitis, biliary obstruction, alcohol toxicity, and a variety of metabolic abnormalities.
The term cirrhosis derives from the Greek word κίρρoς, meaning “tawny,” and was chosen early in the 19th century to describe the gross appearance (tawny, nodular, and firm), and later the microscopic appearance of the chronically diseased, physiologically burned out, and dysfunctional liver. In 1977, an international panel sponsored by the World Health Organization (WHO) defined cirrhosis as “a diffuse process characterized by fibrosis and the conversion of normal liver architecture into structurally abnormal nodules.” The WHO stated that “cirrhosis is a chronic progressive condition that results in liver cell failure and portal hypertension” and observed that vascular abnormalities were an important feature of cirrhosis. These abnormalities included thrombosis, obliteration and recanalization of veins, formation of arteriovenous shunts, and “capillarization” of sinusoids. Etiology was not an important consideration, because there was no cure for cirrhosis. However, reports in the 1980s described the partial “reversal” of cirrhosis, and there is now consistent evidence that effective treatment of the underlying cause may mitigate the fibrotic progression of certain liver diseases, particularly in viral hepatitis.
In 2012, an international group of liver pathologists recommended discontinuing use of the term cirrhosis , substituting instead the diagnostic concept of “an advanced stage of chronic liver disease,” with emphasis on etiologic cause, grade of activity, and whether features are present that suggest progression or regression of the fibrotic process, presence of other diseases, or features indicating risk for malignancy. Liver morphology should be used to help create an integrated assessment of the patient’s clinicopathologic status. This recommendation is in accord with the increasing awareness in the hepatology community concerning the need for pathophysiologic, rather than morphologic, classification of cirrhosis. Indeed, identifying when a liver truly becomes “cirrhotic” is a major diagnostic dilemma, depending on whether one relies on the clinical features of advancing liver disease, imaging modalities, liver biopsy, or an increasing array of noninvasive tests of liver fibrotic status. All this being said, the term cirrhosis has been used for two centuries and will remain in use for the foreseeable future. This chapter is titled and written with that historical premise in mind.
Cirrhosis is defined by the presence of certain anatomic abnormalities of liver structure. However, a presumptive diagnosis of cirrhosis can often be made when certain clinical consequences of cirrhosis are found. These consequences may be mechanical, functional, or neoplastic. The mechanical effects are related to obstruction of blood flow in the liver, which leads to increased pressure in the splanchnic veins and a high risk for rupture of esophageal varices. Obstruction progresses gradually as cirrhosis develops but may worsen suddenly after thrombosis of the portal vein. Obstruction is associated with the opening of multiple intrahepatic and extrahepatic collateral channels that allow shunting of splanchnic blood past the hepatic parenchyma. Hepatic functional deficits are, in part, related to this portosystemic shunting but also to loss of hepatocellular mass and intracellular retention of bile salts and other toxic substances. Many of the clinical effects of cirrhosis are found in other organs. Renal and pulmonary failure may occur secondary to the systemic hemodynamic response to cirrhosis. Hemorrhage may occur because of platelet dysfunction, platelet sequestration in the spleen, and decreased synthesis of proteins of the coagulation cascade. Neoplasia is a not infrequent late occurrence in patients with cirrhosis. The lifetime risk of hepatocellular carcinoma exceeds 50% in patients with some forms of cirrhosis. Cholangiocarcinoma may be a late complication of intrahepatic or extrahepatic biliary disease. Hence, the development and consequences of cirrhosis are of concern in every form of chronic liver disease.
Using the 1977 consensus criteria, cirrhosis is defined anatomically by the presence throughout the liver of fibrous septa that subdivide the parenchyma into nodules ( Fig. 50.1 ). Several features in this definition should be emphasized.
The entire liver must be involved. Occasionally, a focal injury to the liver can cause changes that are histologically similar to cirrhosis—for example, in focal nodular hyperplasia and in liver parenchyma adjacent to neoplasms, abscesses, and other mass lesions. Because these conditions do not have clinical manifestations of cirrhosis, the term focal cirrhosis may be misleading and should be avoided.
The fibrous scarring may be in the form of delicate bands connecting portal tracts and centrilobular terminal hepatic veins in a portal-to-portal, portal-to-venous, or centrilobular venous-to-venous pattern, or they may occur as broad fibrous tracts that obliterate multiple adjacent lobules.
Parenchymal nodules are created by fibrotic isolation of islands of hepatic parenchyma. The regenerative response of hepatocytes may produce nodules with a spherical conformation. However, regeneration and formation of spherical nodules are not required for the definition of cirrhosis; subdivision of the liver by fibrous tissue is required.
Cirrhotic livers may exhibit various patterns that can be attributed to the specific location within the hepatic microvasculature of the etiologic injury, the tempo and duration of the disease, and the presence or absence of an inactive period before histologic sampling. In the last instance, substantial remodeling of fibrous tissue can occur with time and may obscure the architectural pattern of the initial injury.
Although there are many possible types of injuries to the liver, only those with certain characteristics result in cirrhosis. The injury must not be too severe; otherwise, the liver will fail quickly, and the patient will die or undergo liver transplantation before there is sufficient development of fibrosis and architectural remodeling. For example, an acute overdose of acetaminophen causes severe hepatic necrosis and may kill the patient, but it will not produce cirrhosis in those who survive. Submassive hepatic necrosis, with healing, may produce deep scars within the liver that on regeneration of residual liver tissue can lead to hepar lobatum (discussed later). In this condition, the liver is misshapen but, again, not cirrhotic. The process of repeated minor injuries with progressive damage more typically leads to cirrhosis.
The progression of chronic liver disease is highly variable. The point at which a liver becomes cirrhotic is rather subjective and is a frequent source of interobserver disagreement. Establishing a diagnosis of cirrhosis on the basis of percutaneous or transjugular needle biopsy sampling (which samples less than 1/10,000 of the liver mass) can be difficult, particularly if fibrous septa are widely spaced or have regressed. Fortunately, clinical data often provide valuable guidance as to whether any abnormal findings observed in percutaneous liver biopsy tissue are representative of the entire liver. Supporting clinical data include physical examination findings (e.g., ascites, caput medusae, spider angiomas, gynecomastia) and impressions gained from imaging studies or intraoperative visualization of the organ. Laboratory data may not reveal abnormalities, because serum levels of albumin, clotting factors, urea, alkaline phosphatase, aminotransferases, and bilirubin can be normal in a patient who has quiescent cirrhosis with minimal ongoing damage but has not yet experienced hepatic failure, whereas a patient with massive hepatic necrosis and hepatic failure is not cirrhotic despite profound abnormalities in the above serum parameters. Therefore, laboratory data, per se, do not establish a diagnosis of cirrhosis, although various laboratory multiparametric indices have been advanced as capable of providing diagnostic guidance.
Occasionally, a severe focal injury to the liver results in focal histologic changes indistinguishable from cirrhosis on percutaneous needle biopsy; this focal change is not considered true cirrhosis. When this question arises, having definitive information from clinical evaluation and from imaging studies on the general status of the liver, or a biopsy sample from elsewhere in the liver, is critical to determine whether a fibrotic process is focal or diffuse.
The 1977 definition of cirrhosis is used by pathologists to recognize cirrhosis, but the definition does not rely on understanding its pathogenesis. Liver cirrhosis is not, strictly, the end stage of hepatic scarring. Rather, it is a dynamic, biphasic process dominated on the one hand by progressive parenchymal fibrosis and on the other by severe disruption of vascular architecture and distortion of the normal lobular architecture. The main anatomic elements ( Box 50.1 ) include deposition of collagen in the parenchyma and portal tracts, arterialization of parenchymal sinusoids, obliteration of small hepatic and portal veins with resultant loss of hepatocytes through a process referred to as parenchymal extinction , abnormal vascular physiology, and regeneration of hepatocytes. Although the importance of these elements is widely appreciated, there is healthy debate concerning the role of each in the pathogenesis of cirrhosis, particularly because the design of effective interventions to delay or reverse the development of cirrhosis depends on understanding its pathogenesis. The concepts thought to be operative in the genesis of cirrhosis of any cause are summarized here. A discussion of the causes of hepatocellular death is beyond the scope of this chapter.
Deposition of collagen
Arterialization of parenchymal sinusoids
Obliteration of portal veins and hepatic veins
Resorption of fibrous tissue
“Reversal” of cirrhosis
Collagen in the Liver
Collagen accumulation is a prominent feature of cirrhosis. In the normal liver, collagen types I and III are concentrated in the portal tracts and around terminal hepatic veins, with bundles occasionally located between hepatocytes and endothelial cells in the space of Disse. Strands of type IV collagen (reticulin) are present in the space of Disse, where they form a delicate and uniform network that supports the liver cell plates. In cirrhosis, excessive amounts of types I and III collagen are deposited in the portal tracts, along individual liver cell plates in the space of Disse, and in regions of necroinflammatory collapse. A variety of noncollagenous matrix proteins are also deposited in the space of Disse. In cirrhosis, the amounts of collagen, glycoproteins, and proteoglycans can increase severalfold. On a percent area basis, total extracellular matrix components can increase from 5% in normal liver to 25% to 40% in cirrhosis. Some of this is only an apparent increase, because condensation of the normal structural collagen and other matrix components occurs during parenchymal collapse and extinction.
The two main cell types that synthesize collagen in the liver are hepatic stellate cells and portal fibroblasts. Hepatic stellate cells reside in the subendothelial space of Disse in the sinusoidal walls. They are normally distended with lipid droplets containing retinyl esters and other fat-soluble vitamins. During hepatic injury, stellate cells are stimulated by inflammatory mediators to become myofibroblasts: They lose their fat globules, express α-smooth muscle actin in the cytoplasm, and commence proliferation and collagen synthesis. It has recently been discovered that autophagy of the lipid droplets and catabolism of the retinyl esters and triglycerides provide the energy to drive the activation of stellate cells. Experimental inhibition of stellate cell autophagy inhibits their activation.
Stellate cell activation and sinusoidal fibrosis are readily reversible within weeks after cessation of injury. When stellate cells are activated in chronic low-grade disease, the liver cell plates are able to maintain their structure while collagen is deposited in the space of Disse, giving an appearance known as pericellular fibrosis or sinusoidal fibrosis . This type of delicate collagen is most easily appreciated in the perivenular regions (Rappaport zone 3). With time, collagen is deposited along the entire length of the sinusoid. Alternatively, widespread injury to hepatocytes, as in alcoholic hepatitis or some forms of drug injury (e.g., amiodarone), may activate stellate cells throughout the liver, leading to extensive deposition of sinusoidal collagen. In either instance, the total matrix in the space of Disse increases and changes from one that contains delicate interspersed strands of fibrillar collagen (types III and IV) to one composed of a dense matrix of basement membrane–type matrix proteins, which closes the space of Disse to protein exchange between hepatocytes and plasma. In general, abnormal matrix deposition within the space of Disse occurs in those parts of the parenchyma where cell injury and inflammation are greatest.
Portal tract fibroblasts differ from stellate cells in location and physiology. These cells are activated by injury within the portal tracts, particularly in biliary disease, which leads to fibrosis in the region of the ducts and ductules. Peribiliary myofibroblasts are capable of rapid proliferation and deposition of collagen. The epithelial-to-mesenchymal transition of bile duct epithelial cells to a myofibroblast phenotype also appears to contribute to the fibrosis that develops around bile ducts in chronic biliary disease. As a result, fibrosis arising from biliary tract disease can run an aggressive course; an example is complete biliary obstruction in infants with extrahepatic biliary atresia, in whom the liver becomes cirrhotic by 9 weeks of age (see Chapter 54 ). At the opposite end of the spectrum is the exceedingly indolent progression of portal tract fibrosis to cirrhosis in primary biliary cirrhosis, which may extend for 20 or more years (see Chapter 47 ). In either instance, bridging fibrous septa between portal tracts develop throughout the liver, thereby fulfilling the criteria for cirrhosis. A curious feature of biliary-type fibrosis is that the lobular parenchyma is not induced to regenerate substantially until the liver is extensively fibrotic. Hence, biliary-type fibrosis subdivides the liver into a jigsaw-like pattern during its progression, and cirrhosis may be a very late feature in the course of the disease.
Hepatic Arterialization and Capillarization
Arterialization of the liver in cirrhosis has been known for more than a century. In 1907, Herrick perfused cadaver livers and demonstrated that resistance to flow in the hepatic artery of cirrhotic livers was markedly decreased. This fact is documented daily by ultrasonographers when they examine patients with cirrhotic livers and find increased arterial flow in the liver, along with sluggish or even reversed flow in the portal vein. Moschcowitz first used the term capillarization to describe the light microscopic appearance of arterialization in the cirrhotic liver as a granulation tissue response—that is, arterial growth into inflamed tissue. Schaffner and Popper used the term capillarization to represent a constellation of ultrastructural changes, including a decrease in the number and size of sinusoidal endothelial fenestrations, loss of hepatocellular microvilli, and an increase in basement membrane material. By this definition, documentation of capillarization requires electron microscopy. Specifically, in the normal liver, sinusoidal endothelial cells lack a basement membrane and exhibit fenestrations approximately 100 nm in diameter, occupying between 2% to 3% of the area of the endothelial cell. Deposition of extracellular matrix in the space of Disse is accompanied by loss of fenestrations in the sinusoidal endothelial cells. With the development of cirrhosis, the diameter of the fenestrations slightly decreases but the area occupancy (porosity) falls to less than 0.5%.
The sinusoidal endothelium in the cirrhotic liver may express CD34. Because this expression is a normal property of arterial endothelium, CD34 positivity is considered a marker of arterialization of the sinusoids. Sinusoidal arterialization is accompanied by α-smooth muscle actin staining in the sinusoidal wall ( Fig. 50.2 ), reflecting the accompanying transformation of hepatic stellate cells into myofibroblasts. It appears that capillarization leads to loss of normal sinusoidal endothelial cell generation of nitric oxide (NO). This loss of NO generation creates a microenvironment that is permissive for stellate cell activation. Transformation of the stellate cells then increases sinusoidal vascular resistance by tonic contraction of these “myofibroblasts.” Sinusoidal fibrosis in the perivenular region of the lobule may also partially obstruct vascular outflow, creating postsinusoidal vascular resistance.
Transformation of sinusoidal vascular channels is widely considered to be an explanation for functional deficits in blood-hepatocyte solute exchange. Rapid flow in sinusoids may represent an effective arteriovenous shunt, resulting in a further decrease in solute exchange. Although these effects may decrease solute transport into hepatocytes, capillarization or arterialization may also be viewed as a protective form of adaptation that allows the hepatocytes to survive in a high-pressure, high-flow environment.
Arterialization also occurs at the level of portal tracts, where an increased number (and size) of arterial profiles is seen in a variety of conditions, including cirrhosis and use of oral contraceptives. Arterialization of small portal tracts in cirrhosis is usually accompanied by obliteration of adjacent portal veins. Portal vein loss may result from portal tract inflammation in chronic hepatitis or from congestive changes (congestive portopathy) after hepatic venous outflow is compromised (discussed in the next section). Obliteration of small portal veins increases presinusoidal vascular resistance for blood inflow via the splanchnic system. Resistance to hepatic arterial blood flow decreases, owing to an increased arterial capacity, whereas resistance to portal vein blood inflow increases. Hepatic arterial blood pressure is sufficient to supply blood to the liver, but the low pressure of the splanchnic system is not able to overcome the pressure impedance, leading to portal hypertension.
Parenchymal extinction is defined as a focal loss of contiguous hepatocytes ( Fig. 50.3 ). Detailed morphologic studies by Wanless and colleagues demonstrated that parenchymal extinction plays a critical role in the pathogenesis of cirrhosis.
Hepatocyte apoptosis and necrosis occur in all types of liver diseases that progress to cirrhosis. The mechanisms are diverse and include lymphocyte-mediated injury, rupture of triglyceride-laden hepatocytes, bile-salt toxicity, and various metabolic stresses. Most of these injuries, if accompanied by low-grade spotty necrosis or apoptosis, lead to local replacement and complete healing. Progressive disease occurs when these injuries are accompanied by a stromal reaction that includes deposition of extracellular matrix, increased sinusoidal vascular resistance, and obstruction of blood flow. The convergence of these injuries leads to contiguous loss of hepatocytes (see Fig. 50.3 ). These extinction lesions may involve a small portion of an acinus, larger units of one or more adjacent acini, or even a whole lobule. The contiguous cell loss is ultimately the result of focal ischemia resulting from obstruction of veins or sinusoids. Naturally, the size of extinction lesions depends on the size of the obstructed vessels.
The concept of parenchymal extinction is important because it incorporates the following perceptions: (1) parenchymal extinction is not directly caused by the initial hepatocellular injury but is an epiphenomenon caused by innocent bystander injury of the local vessels; (2) each parenchymal extinction lesion (PEL) has its own natural history and may be in an early or late stage of healing; (3) cirrhosis develops simultaneously with the accumulation of numerous independent and discrete PELs throughout the liver; and (4) the form of cirrhosis is largely determined by the distribution of the vascular injury. Importantly, parenchymal extinction may progress long after cirrhosis is already established, leading to slow conversion of a marginally functional liver into a sclerotic organ incapable of sustaining life.
The pathogenesis of vascular obstruction depends on the size of the vessels and is outlined in Figure 50.4 (see also Fig. 50.3 ). Most small-vessel obliteration is secondary to local inflammation. Although thrombosis may be important in veins of all sizes, it is especially important for blockage of large veins. Most PELs are produced by blockage of veins larger than 100 µm in diameter, because obstruction at this site cannot be easily circumvented by collateral flow within the sinusoids. Obstruction of several adjacent sinusoids is also difficult to circumvent. This raises the issue of congestion, which occurs whenever blood entering the vasculature exceeds the ability of the outflow tract to carry that blood, a state known as in-out imbalance . Congestion is particularly severe when there is total obstruction of the outflow tract or when there is increased inflow in the presence of partial outflow obstruction. Congestive injury is exacerbated by reactive hyperemia, shunt formation, and angiogenesis. If inflow is marked, congestion occurs even when the tissue has normal outflow capacity.
Although the obstructive vascular events leading to parenchymal extinction play out largely in blood vessels larger than 100 µm in diameter, activation of the coagulation cascade within sinusoids releases thrombin, which can interact with the proteinase receptor PAR1 on hepatic stellate cells. This is a stimulus for stellate cell activation and transformation to myofibroblasts, which further contributes to the increased vascular resistance and congestive pathophysiology described earlier.
Therefore, PELs are the result of local failure of the microvasculature, usually because of obstruction of hepatic veins or sinusoids. The mechanism for formation of PELs is detailed in Figure 50.5 (see also Fig. 50.4 ).
In early chronic liver disease, PELs are recognized by the close approximation of the terminal hepatic vein and the adjacent portal tract (see Figs. 50.3 and 50.5 ). They may be difficult to recognize, because damaged small hepatic veins only appear as a few collagen bundles lying adjacent to a portal tract. PELs become more evident as they aggregate and involve larger and more easily recognizable hepatic veins. PEL aggregates may be evident as two or more portal tracts that are bound together with a hepatic vein remnant apparent in the intervening space (see Fig. 50.2 ). With progression of disease, progressive obstruction of hepatic veins and secondary arterial dilatation cause further congestive injury in the tissue located between the lesions. This creates a self-perpetuating pathophysiology that may eventually lead to interconnected portal tracts throughout the whole liver.
When a region of parenchyma becomes extinct, it collapses so that a portal tract becomes closely associated with an adjacent terminal hepatic vein. This close approximation offers an opportunity for the artery in the portal tract to drain directly into the collapsed perivenous tissue. Often, these arteries can be seen supplying a pool of blood surrounded by atrophic hepatocytes. In older lesions, there is a well-demarcated blood-filled channel; this suggests a stable, high-flow and high-pressure conduit connecting a small artery to a small hepatic vein. This appearance has been interpreted as an arteriovenous shunt.
After parenchymal extinction, the formation of bona fide bridging fibrous septa between portal tracts and terminal hepatic veins enables portovenous and arteriovenous shunting through de novo vascular channels, effectively bypassing the parenchymal nodules. Shunted blood flow through these “fast” vascular channels leaves the remainder of the hepatic parenchyma almost bereft of meaningful blood flow. This progression also helps explain the increased blood flow observed in sinusoids of the cirrhotic liver in the midst of relative underperfusion of the liver parenchyma as a whole. A remarkable fraction of nutritive blood flow may pass through these intrahepatic functional shunts, contributing to ongoing hepatocellular necrosis and the further generation of PELs. Unfortunately, compression of the shunt channels by contiguous regenerating nodules maintains an increased transhepatic vascular resistance.
Vascular injury in cirrhosis can be divided into an early primary phase, a later congestive phase, and, ultimately, vascular thrombosis. In the primary phase, venous and sinusoidal obstruction is caused by local inflammation occurring in the course of chronic hepatitis ( Fig. 50.6 ). The generation of soluble proinflammatory mediators in this setting is a powerful stimulus of fibrogenesis. However, one of the most powerful stimuli of fibrogenesis in many organs is the organization of exudates, especially those rich in fibrin. In the chronically congested liver, the development of capsular fibrosis attests to the presence of this fibrogenic mechanism ( Fig. 50.7 ). This is the basis for the schematic diagram in Figure 50.8 , which indicates that instances of fibrosis can be divided into those stimulated by conventional inflammation and those stimulated by chronic edema and exudation. These two pathways may converge to generate the pattern of fibrogenesis typically seen in chronically injured livers.
Collagen accumulation is determined by the rates of collagen synthesis and resorption ( Fig. 50.9 ; see also Fig. 50.8 ). Most fibrosis likely accumulates late, when congestive forces cause interstitial exudate and collagen deposition that exceeds the resorptive capacity of the liver. In this formulation, obstructed hepatic veins and PELs occur before septa and before fibrosis; therefore, these lesions are at the leading edge of the pathogenesis of cirrhosis. The preexisting structural collagen of the liver condenses during the formation of PELs and is incorporated into septa. Although this is not true fibrosis, the presence of such collagen is a striking feature of these lesions.
In the later congestive phase of liver injury, a self-perpetuating pattern of progressive liver injury is created. Specifically, when there is in-out imbalance of blood flow, the high transmural pressure leads to edema, hemorrhage, and narrowing of venous and sinusoidal lumina, followed by intimal fibrous thickening of these vessels. This injury, known as congestive hepatopathy , causes a decrease in the outflow capacity of the tissue that worsens obstruction and, therefore, congestion. Thus, congestive hepatopathy establishes a positive feedback loop of progressive tissue injury.
Thrombosis is an important insult. In angiographic and ultrasonographic studies, portal vein thrombosis has been detected in 0.6% to 16.6% of cirrhotic patients, and grossly visible portal vein fibrosis or thrombosis has been found in 39% of cirrhotic livers at autopsy. Veno-occlusive lesions of hepatic veins smaller than 0.2 mm in diameter have been found in as many as 74% of cirrhotic livers examined at autopsy. Obliterative lesions are found in 36% of portal veins and 70% of hepatic veins in livers removed at liver transplantation. The distribution of obliterative lesions is more uniform in portal veins than in hepatic veins, consistent with the concept of propagation of multifocal thrombi downstream from their site of origin. Portal vein lesions are associated with prominent regional variation in the size of cirrhotic nodules. Hepatic vein lesions are associated with regions of confluent fibrosis and parenchymal extinction. The compelling conclusion is that thrombosis of medium- and large-sized portal veins and hepatic veins is a common occurrence in cirrhosis and may represent a final common pathway for the propagation of parenchymal extinction to full-blown cirrhosis. Furthermore, cirrhotic livers are susceptible to thrombosis because of sluggish or reversed blood flow and the prothrombotic effects of sepsis and cholestasis, thus creating an opportunity for continued loss of functional residual liver parenchyma.
After childhood, the normal liver becomes a stable organ with slow turnover of hepatocytes. However, on injury or surgical reduction, the liver cells proliferate. Normal human liver can restore approximately three fourths of its own mass within 6 months. Hepatocytes, bile duct epithelial cells, and hepatic progenitor or stem cells maintain the potential to multiply during adult life. Depending on the severity of the primary injury, liver regeneration may occur by at least two mechanisms. In brief, with mild to moderate hepatocellular loss, mature hepatocytes undergo replication. More extensive or massive hepatic necrosis stimulates proliferation of progenitor cells within the periportal region, particularly when necroinflammation occurs at the portal-parenchymal interface. Proliferation of these cells gives rise first to so-called ductular hepatocytes, in which ductular structures containing cuboidal cells and slightly larger cells with mitochondria-rich cytoplasm are present. With time, these cells mature into definitive hepatocytes or cholangiocytes and may repopulate damaged parenchymal regions and bile duct structures, respectively.
Parenchymal regeneration is recognized initially by the twinning of liver cell plates, evident as a double line of hepatocytes with nuclei seemingly running in parallel. Twinning of cell plates may remain for some months after regeneration before new sinusoidal channels develop and the nuclear alignment dissipates. If regeneration is recent, the hepatocytes lack lipofuscin, because this pigment accumulates with time in the normal liver. Regeneration also is characterized by increased numbers of binucleate or multinucleate hepatocytes, reflecting replication of nuclear material. Hepatocyte nuclei may be more uniform in size, because anisonucleosis increases with patient age in the normal liver and may not be as evident in regenerating liver. Finally, regeneration imparts a rounded appearance to the expanding contours of residual parenchyma, which is demonstrated with a reticulin stain. To the extent that fibrosis and cell death precede hepatocellular regeneration, a residual viable parenchymal island may stand out in the midst of necroinflammation and developing fibrosis.
It is not until hepatocellular regeneration occurs that the characteristic nodular transformation of cirrhosis becomes manifest. On thickening of the liver cell plates, the parenchyma expands against the constraining fibrous septa and acquires a spherical shape. The hepatocyte plates abutting the fibrous septa become compressed and are bent outward by the less-constrained plates toward the interior of the nodule. Poor regeneration or failure to keep up with the pace of collapse results in variants of cirrhosis in which the feature of rounded contours may not be prominent.
The ultimate size of the nodule is determined, in part, by the anatomic location of the antecedent fibrous septa. If matrix deposition occurs at the acinar level, the resulting nodules will grow out of monoacinar units and will be small. If matrix deposition encompasses many acinar units (multiacinar), the growing nodules may be much larger and will retain components of the preexisting acini, including intact portal tracts.
In some cases of cirrhosis, vast expanses of bile ductules within the fibrous septa coexist with the interspersed hepatocellular nodules. These ductules may occur in cirrhosis of almost any cause and are not necessarily the result of biliary obstruction. Hyperplasia of ductules is associated with lengthening and increased tortuosity of existing channels and with extensive sprouting of new channels. This change is reminiscent of the massive proliferation of ductular structures within the hepatic parenchyma, or at the interface between parenchyma and portal tracts, that occurs in massive hepatic necrosis and implicates a proliferation of periportal progenitor cells. With time, these ductular structures may mature into hepatocellular parenchyma or bile ductules. Therefore, the presence of expanses of bile ductules in a cirrhotic liver points toward episodes in the recent past in which there was extensive parenchymal destruction; the ductules represent an intermediate stage of a massive regenerative response.
Natural History and Reversibility of Cirrhosis
Cirrhosis is traditionally viewed as an end stage in the evolution of many types of chronic liver diseases. However, clinical reports have indicated that on cessation of the injurious process, cirrhosis may reverse, or at least improve, histologically. Tissue samples from some patients with established cirrhosis have revealed incomplete septal cirrhosis or apparent absence of fibrosis after successful treatment. This evolution has been documented in patients with hemochromatosis, autoimmune hepatitis, Wilson disease, primary biliary cirrhosis, schistosomiasis, extrahepatic biliary obstruction, alcoholic disease, chronic viral hepatitis B or C, or postjejunal bypass steatohepatitis. The concept of reversibility of fibrosis, potentially including patients with a histologic or clinical diagnosis of cirrhosis, is now not only widely accepted but a hoped-for outcome after pharmaceutical treatment of some forms of chronic liver disease, particularly viral hepatitis.
In many different experimental models of cirrhosis reversal, collagen is resorbed within weeks after cessation of injury. The mechanism of resorption of fibrous extracellular matrix involves activation of tissue metalloproteinases. Likewise, stellate cells and activated portal tract fibroblasts may undergo apoptosis and subside. Fibrosis may even disappear in cases of long-term quiescent cirrhosis. In this situation, the diagnosis of cirrhosis may be made by demonstrating severe paucity of small hepatic and portal veins, even when fibrous septa are highly regressed and not well represented within a biopsy specimen. Indeed, a revised definition of cirrhosis can be made wherein it represents a condition of widespread obliteration of small hepatic veins (i.e., venopenia) . This definition is sufficient because if there is severe venopenia, all other histologic features of cirrhosis will ultimately follow.
Therefore, the histologic appearance of cirrhosis depends on the age of accumulated tissue damage and the time of dormancy with an opportunity to resorb fibrous tissue. The presence of PELs is a helpful indicator. If the causative injury is long past, only old PELs will be present. If the causative injury is long past, the liver will contain lesions only in the late stages of repair. New extinction lesions are easily seen in livers with moderate to severe activity. They are recognized as areas of bridging necrosis or of focal intense congestion with atrophy and clusters of apoptotic cells. More commonly, especially in low-grade chronic hepatitis, the lesions are recognized as subtle atrophy, sinusoidal dilatation, clusters of apoptotic cells, and approximation of hepatic veins close to portal tracts.
Old extinction lesions predominate in livers in which the primary disease has remitted, either spontaneously or after successful treatment. Fibrosis tissue is progressively removed from extinction lesions so that broad septa become delicate and delicate septa become incomplete (“perforated”) or disappear. Thus, micronodular cirrhosis may remodel to macronodular cirrhosis, incomplete septal cirrhosis, or near-normal liver (and be discovered only if there is “noncirrhotic” portal hypertension). The problem of diagnosing regression of cirrhosis is discussed later.
Even with substantial resorption of fibrous septa, restoration of the hepatic architecture to a normal state probably does not occur. Limiting factors are the persistence of vascular abnormalities—that is, outflow obstruction and arterialization ( Figs. 50.10 and 50.11 ). If these vascular factors are sufficient to cause continued hepatocellular injury or interstitial exudation, septa will not resorb, and some degree of cirrhosis or incomplete septal cirrhosis will remain. Arterialization is important because even if hepatic vein outflow is not limiting, elevated sinusoidal pressure prevents the regeneration of obliterated small portal veins so that noncirrhotic portal hypertension may remain. Moreover, residual sclerosis in portal tracts may still lead to persistent presinusoidal resistance to splanchnic blood flow, leading to continued clinical evidence of portal hypertension.
The previous description indicates that the equilibrium of injury and repair depends on the local balance of inflow and outflow of blood and on the continued presence (or absence) of the proinflammatory disease environment. This equilibrium is continually affected by activity of the primary disease, congestive hepatopathy, persistence of fibrosis in anatomically strategic locations, and catastrophic events such as portal or hepatic venous obstruction by tumor or thrombosis.
Anatomic Classification and Pathology
The main macroscopic types of cirrhosis are referred to as micronodular and macronodular . The former consists of nodules mostly smaller than 3 mm in diameter, the latter of nodules mostly larger than 3 mm ( Box 50.2 ). The term mixed cirrhosis describes livers in which nodules both larger and smaller than 3 mm coexist. The definitions of macronodular and micronodular cirrhosis have been modified over the years, from an original cutoff value of 10 mm (before 1976) to the current WHO cutoff of 3 mm. This modification was intended to help distinguish alcoholic cirrhosis from other etiologic types. However, since the WHO criteria appeared, hepatitis C cirrhosis has been shown to be mostly micronodular, and alcoholic micronodular cirrhosis has been shown to transform to a more macronodular form during inactive disease. In addition, the progressive stages of biliary cirrhosis, from incomplete, to macronodular, mixed, and finally micronodular forms of cirrhosis, have been described. Therefore, this anatomic classification is not adequate to allow a precise correlation with etiology. Rather, it is clear that the severity of activity and duration of inactivity are important. The evolutionary nature of these anatomic types is shown in Figures 50.12 and 50.13 .
Micronodular cirrhosis with regional parenchymal extinction
Mixed micronodular/macronodular cirrhosis
Postnecrotic cirrhosis after subacute massive necrosis
Noncirrhotic portal hypertension
Incomplete septal cirrhosis
Nodular regenerative hyperplasia (see Chapter 51 )
Hepatoportal sclerosis (see Chapter 51 )
Congenital hepatic fibrosis (see Chapter 54 )
Despite these caveats, certain generalizations can be made. Micronodular cirrhosis is associated with diseases in which there is uniform injury to all acinar units. The result is that most small hepatic veins are damaged, parenchymal extinction extends to involve all portal tracts, and subsequent regenerative nodules are small and devoid of portal tracts. In contrast, livers with macronodular cirrhosis are irregular, with less severe destruction of hepatic veins, so that most regenerative nodules contain some portal tracts and intact hepatic veins. Therefore, based on the single parameter of hepatic vein destruction, micronodular cirrhosis is more severe than macronodular cirrhosis.
In the early evolution of micronodular cirrhosis, the overall shape and external appearance of the liver may not be greatly altered. The organ weight may be normal or increased, and the left lobe may enlarge disproportionately. Parenchymal nodules may be difficult to discern by observation of the capsule of the liver. With time, organ weight decreases, and the capsular surface may become studded with a myriad of small protruding nodules between the shallow and regular fibrous indentations that define the nodules. On cut section, the nodules are small and uniform but may be difficult to see without slight magnification. As the cirrhotic liver becomes small, the liver becomes progressively firm to the touch, and septa are more wide and prominent. In severe examples, residual small islands of parenchyma appear to float in a sea of fibrous tissue.
Alcohol abuse is the most frequent cause of micronodular cirrhosis in Europe and North America. However, increasingly common is the cirrhotic outcome of nonalcoholic fatty liver disease. Hepatitis C viral infection, being less prone to bouts of severe necroinflammatory injury, progresses with time from an initial macronodular pattern toward a more micronodular pattern of cirrhosis (unlike hepatitis B viral infection, which remains macronodular). Less common causes for the micronodular pattern of cirrhosis include hemochromatosis, chronic biliary obstruction caused by primary biliary cirrhosis or primary sclerosing cholangitis, drug toxicity, and many types of metabolic diseases of infancy and childhood (see Chapter 54 ). With chronic extrahepatic venous outflow obstruction (see Chapter 51 ), the liver also is finely subdivided by fibrous tissue (see Congestive Cirrhosis ).
In alcoholic liver disease, the liver may show a diffuse tan or yellow color caused by severe fatty change or, in the setting dominated by alcoholic hepatitis, it may be a characteristic red color. In hemochromatosis, heavy iron deposition imparts a dark reddish-brown color to the liver. Destructive biliary diseases such as primary biliary cirrhosis or primary sclerosing cholangitis, while rendering the liver a deep yellow or green color, impart to the liver capsule a progressive fine septal scarring pattern resembling pigskin. Although any decompensated liver may become yellow or green, it is seldom as deeply pigmented as in patients with a biliary disease. There are rarely any distinguishing macroscopic features in micronodular cirrhosis caused by drug toxicity.
The condition known as cirrhosis with regional parenchymal extinction (also known as confluent hepatic fibrosis ) occurs when macroscopic expanses of liver tissue become purely fibrotic ( Fig. 50.14 ). This appears to be the natural progression of cirrhosis when there is progressive regional vascular compromise. Regional parenchymal extinction may be 1 to 2 cm in greatest dimension or even much larger. The capsular surface overlying large areas of extinction may be finely granular and wrinkled, rendering a pigskin appearance.