CHAPTER 5 Biliary Disease
There are many sites along the biliary tree where bile flow may be interrupted, from the bile canaliculi and smallest intrahepatic ducts to the large bile ducts and duodenum (Fig. 5.1). Damage or obstruction at these various sites may result in visible bile in histological sections (cholestasis), altered bile-duct morphology, changes within the portal tracts and periportal parenchyma, or combinations of these. Diseases of the larger ducts must be distinguished from diffuse intrahepatic diseases because of different clinical management, and liver biopsy is often helpful in this respect. However, diseases of large bile ducts outside and within the liver share pathological features and may be amenable to similar forms of treatment; for this reason the term ‘extrahepatic biliary obstruction’ is not used in this chapter. Carcinoma of the main hepatic ducts, for example, may be situated wholly within the liver, yet lead to the changes of large-duct obstruction. This chapter discusses these changes as well as the pathology of primary biliary cirrhosis and primary sclerosing cholangitis, the diagnostic problem of overlap with autoimmune hepatitis and several bile-duct paucity disorders.
Figure 5.1 The biliary tree.
The ramifying structures of the biliary system are shown schematically. The large segmental and area ducts have peribiliary glands (PGs). The finer branches are shown in an enlargement at upper left. Boxes at right show examples of biliary disease at the specific levels affected. PBC = primary biliary cirrhosis; PSC = primary sclerosing cholangitis.
The term cholestasis in clinical and pathological usage refers to impairment of bile flow. Under the light microscope, cholestasis (sometimes called bilirubinostasis) is defined as the presence of bile pigment within bile canaliculi, hepatocytes and other sites. It is the morphological correlate of clinical jaundice. Cholestasis is an important finding in large bile-duct obstruction or in extensive intrahepatic bile-duct disease, but may also accompany the parenchymal damage in certain types of hepatitis. Pure (bland) cholestasis as an isolated lesion requires consideration of several possible aetiologies (Table 5.1), which may not be distinguishable by light microscopy alone. For example, in neonatal and childhood jaundice, cholestasis may result from mutations in bile-salt transport proteins on the canalicular membrane1 or from mitochondriopathies,2 problems discussed further in Chapter 13. In adults, drug hepatotoxicity, circulating endotoxin in septicaemia3 and cytokine release from extrahepatic lymphoma4 are further examples of functional disorders of bile secretory physiology that may lead to intrahepatic cholestasis (discussed further in Ch. 4). The pathologist’s first priority when cholestasis is present, nevertheless, is careful examination of the portal tracts for possible changes of mechanical large bile-duct obstruction, which are described below.
|Bile-salt transporter mutations (e.g. Byler disease)|
|Mitochondriopathies (e.g. Navajo neurohepatopathy)|
|Early large bile-duct obstruction|
Biopsies from patients with large-duct obstruction are much less often seen than formerly because of improved imaging methods. However, the pathologist needs to be able to recognise the characteristic changes, especially following liver transplantation. From the first weeks of obstruction there is cholestasis in perivenular areas; that is to say, bile is visible under the microscope in the form of bile thrombi (bile plugs) in canaliculi and as yellow-brown pigment in hepatocytes and Kupffer cells (Fig. 5.2). The presence of canalicular bile thrombi distinguishes cholestasis from other pigmentations (see Table 4.1). Kupffer cells in cholestatic areas are enlarged and pigmented, containing both bile and diastase-resistant periodic acid–Schiff (PAS)-positive material. In recovering obstruction the Kupffer-cell changes persist while bile thrombi become smaller and less numerous. Finally, as in residual acute hepatitis, a few diastase–PAS-positive Kupffer cells may provide the only histological evidence of a recent episode of jaundice.
Figure 5.2 Cholestasis.
At first the hepatocytes in areas of cholestasis show little change, but with time they often become swollen. Their nuclei increase in size and number and a few apoptotic bodies and mitoses may be seen, indicating increased cell turnover. Individual hepatocytes or small groups of cells undergo feathery degeneration, characterised by rarefied and reticular cytoplasm (Fig. 5.3). The lesion is focal and the affected cells are typically surrounded by more or less normal hepatocytes. Feathery degeneration may be difficult to distinguish from the ballooning degeneration of hepatitis (see Fig. 6.2) or following liver transplantation, but in ballooning the cytoplasm is often granular rather than feathery and the lesion is more widespread in the lobule.
Figure 5.3 Cholestasis.
In a minority of patients with obstructed ducts bile infarcts form (Fig. 5.4). These are substantial areas of hepatocellular degeneration or death containing pale or bile-stained hepatocytes or discrete rounded cells that are difficult to distinguish from macrophages. There are variable amounts of bile and fibrin, the latter often abundant. Reticulin fibres become progressively more difficult to demonstrate. Bile eventually leaches out of the infarct to leave a barely pigmented and scarcely stained lesion containing the ghosts of hepatocytes. Small bile infarcts may be found in severe cholestasis from any cause; larger infarcts such as the one shown in Fig. 5.4, especially if adjacent to a portal tract, are highly suggestive of bile-duct obstruction. However, because such infarcts are seen in only a minority of patients with obstructed ducts, the diagnosis must usually be established by other criteria.
As a result of these various forms of hepatocellular damage in biliary obstruction, and indeed in cholestasis generally, a certain amount of inflammatory infiltration of the parenchyma is commonly seen after a period of some weeks. This infiltration is usually mild and restricted to the cholestatic areas, unlike the inflammation of an acute hepatitis. When cholestasis resulting from duct obstruction is prolonged, especially in older patients, inflammation and liver-cell damage are occasionally severe enough to raise the alternative possibility of an acute hepatitis. It is then helpful to note that in bile-duct obstruction the liver-cell plates remain for the most part intact, whereas in hepatitis they become irregular as a result of cell loss, swelling and regeneration. Central–portal (zone 3) bridging necrosis is not a feature of biliary obstruction.
Within a few days or weeks of the onset of duct obstruction a characteristic triad of portal changes develops,5 consisting of portal oedema and swelling (Fig. 5.5), infiltration by inflammatory cells, and increased numbers of bile-duct profiles at the margins of the portal tracts (Figs 5.6, 5.7). These marginal bile-duct structures are the most consistent finding in the portal tracts and are rarely absent.5 They may originate from canals of Hering, periportal stem cells or other sources6 and are an early response to the increased portal tract pressure due to obstruction, circulating mediators7 and expression of developmental proteins such as Notch receptors and Jagged proteins.8 The term ductular reaction refers to these proliferated bile ductules accompanied by inflammation and stromal changes at the edges of the portal tracts.9 Usage of ‘ductular reaction’ is now preferable to ‘bile ductular proliferation’ or ‘typical’ and ‘atypical’ bile ductules, which embody considerable imprecision.6 The ductular structures may be of normal calibre or dilated, but are often flattened with small or imperceptible lumens (Fig. 5.7) and variations in nuclear size, staining and location. These structures can be highlighted by immunostaining for cytokeratin 7 or 19 (see Figs 5.24, 5.25). Surprisingly, bile is not usually seen within dilated ducts or ductules in uncomplicated obstruction; when it is present, sepsis should be suspected. The differentiation of the ductular reaction of biliary obstruction from that of chronic liver disease has already been discussed in Chapter 4.
Figure 5.5 Large bile-duct obstruction.
Figure 5.6 Large bile-duct obstruction.
The upper left-hand portion of Fig. 5.6 is shown at higher magnification. The irregular ductular structures at the edge of the portal tract show compressed, narrow lumens and an associated neutrophil infiltrate. (Needle biopsy, H&E.)
Within the oedematous, swollen portal tracts, especially around proliferated bile ducts, an inflammatory infiltrate develops, mediated by the complex interactions of cytokines and cellular adhesion molecules (some produced by biliary epithelium itself10) and proinflammatory agents such as endotoxin.3 Neutrophils are prominent owing to the expression of the chemoattractant interleukin-8 by the ductular cells.11 There may also be other cells including lymphocytes and eosinophils. The presence of a few eosinophils is therefore not in itself sufficient evidence for a diagnosis of drug jaundice. As a result of the proliferative and inflammatory changes of bile-duct obstruction, the outlines of the portal tracts become irregular and the limiting plates of hepatocytes are disrupted to a variable extent. This disruption should be distinguished from interface hepatitis in which the infiltrate is predominantly composed of lymphocytes and plasma cells, and in which the acute inflammatory changes of bile-duct obstruction are not seen.
In a few patients with bile-duct obstruction the portal changes are inconspicuous (Fig. 5.5) or even absent. Biliary obstruction should therefore be considered in the differential diagnosis of canalicular cholestasis without portal reaction (so-called ‘pure’ or ‘bland’ cholestasis). Conversely, portal changes resembling those of duct obstruction are occasionally found in severe acute hepatitis, when the parenchymal alterations make the diagnosis clear. Sometimes similar portal changes are seen without cholestasis near space-occupying lesions such as metastases,12 usually together with sinusoidal dilatation. Portal inflammation without cholestasis is also found in patients with disease affecting one or other part of the biliary tree but without current obstruction of the segment biopsied. It is seen in chronic pancreatitis13 and in patients with acute cholecystitis or choledocholithiasis.14 Biopsies showing only an increased number of well-differentiated bile ductules at the portal interface, unaccompanied by inflammation or stromal changes, have been noted in patients with idiopathic isolated ductular hyperplasia (IDH)15 (Fig. 5.8). These patients have long-standing abnormalities in serum alanine aminotransferase and/or γ-glutamyl transferase, no proven biliary tract disease and an apparently good prognosis (although the cause of this reactive lesion is uncertain).
Figure 5.8 Isolated ductular hyperplasia.
Well-differentiated hyperplastic bile ductules are present in this biopsy from a patient with persistent liver function test abnormalities unrelated to viral hepatitis or demonstrable biliary disease. The absence of oedema and inflammation help distinguish this lesion from large bile-duct obstruction. (Needle biopsy, H&E.) Inset: Immunostain for cytokeratin 7 on a deeper level highlights the bile ductules. (Needle biopsy, specific immunoperoxidase.)
(Case kindly provided by Drs Bernard Traub and John B. Herrington, III, Sleepy Hollow, New York, NY, USA.)
In a few instances of biliary obstruction, bile escapes from a duct into the connective tissue of a portal tract, giving rise to a bile extravasate. This leads to a phagocytic reaction, with or without foreign-body giant cells (Fig. 5.9). Bile extravasates, like large bile infarcts, are almost diagnostic of obstruction but are seen in only a minority of patients. If the extravasate extends beyond the confines of a portal tract into the adjacent parenchyma, the appearances at the periphery of the lesion are very like those of a bile infarct.
Figure 5.9 Large bile-duct obstruction.
When bile duct obstruction persists, the acute inflammatory reaction in the portal tracts is followed by increasing fibrosis. Production of fibrogenic cytokines by bile-duct epithelium contributes to this process.16 Eventually the tracts are linked by broad fibrous septa. There is a variable degree of acute and chronic inflammatory infiltration, the chronic element less striking than in primary biliary cirrhosis. In some patients the lesion appears to progress more by cholangitis than by obstruction, and cholestasis is therefore not always prominent or even present.
Interference with normal secretion of bile leads to several changes in hepatocytes adjacent to portal tracts and fibrous septa. The cells become swollen and separated by fibrous tissue, inflammatory cells and ductular structures (neocholangioles) derived from hepatocytes or bipotential stem cells.17 Their cytoplasm is rarefied and may contain visible bile pigment, Mallory bodies, copper and copper-associated protein (Fig. 5.10). The last is seen in the form of fine red granules on haematoxylin and eosin (H&E) (see Fig. 5.27 Inset), staining variably with diastase–PAS and strongly with orcein or Victoria blue. The combination of all these changes is known as chronic cholestasis or cholate stasis (pseudoxanthomatous change, precholestasis) on the basis that some of the alterations probably result from the accumulation of toxic bile salts. Canalicular cholestasis is sometimes seen between the affected hepatocytes. The hepatocellular changes, ductular proliferation and associated fibrosis in the periportal or periseptal region in effect produce an irregular interface with the parenchyma.18
Figure 5.10 Chronic cholestasis.
The fibrous septa which eventually form in chronic biliary tract disease surround and outline groups of classical hepatic lobules, leaving the normal vascular relationships essentially intact. Islands of parenchyma with characteristic protruding studs resemble the pieces of a jigsaw puzzle or landmasses on a map (Fig. 5.11). Spherical nodules are sparse at first, in spite of evidence of liver-cell hyperplasia in the form of thickened liver-cell plates, seen particularly in patients with associated portal hypertension.19 An occasional rounded parenchymal island may merely represent a tangential section of a complex parenchymal mass such as the one shown in Fig. 5.11, rather than a true regeneration nodule of cirrhosis. This is especially common just deep to the liver capsule. A histological diagnosis of cirrhosis should therefore be made with caution, because at a fibrotic, pre-cirrhotic stage considerable resolution can occasionally result if an obstruction is relieved.20 Eventually, true secondary biliary cirrhosis develops, its biliary origin still evident from nodule shape and the regular, broad fibrous septa composed of loose collagen bundles with parallel arrangement (Fig. 5.12). A zone of oedema containing proliferated ductules is often diagnostically helpful and may be striking even at low magnification (the ‘halo effect’) (Fig. 5.11). Thus, many different structural characteristics make it possible to diagnose chronic biliary tract disease, even in the absence of cholestasis. Finally, however, an end-stage cirrhosis forms, no longer necessarily recognisable as biliary in origin.
Figure 5.11 Secondary biliary cirrhosis.
In biliary obstruction the inflammatory infiltrate around bile ducts in small portal tracts typically includes neutrophils. There is, therefore, cholangitis in a strictly histological sense, but this does not imply that there must be bacterial infection of the biliary tree or clinical ascending cholangitis. In the latter, neutrophils are more numerous and are found not only around ducts but also in their walls and lumens21 (Fig. 5.13). Paradoxically, interlobular bile ducts are most affected, and larger ducts may appear histologically normal. The wall of a small duct may rupture, leading to abscess formation in the portal tract. Neutrophils are seen in the sinusoids and abscesses may form in the acini. Associated lesions include fibrin thrombi in portal-vein branches, pyelophlebitis and various degrees of parenchymal necrosis,22 the last probably related to hypoperfusion of the parenchyma. Cholestasis is more often absent than present. Causes of ascending cholangitis include cholecystitis and choledocholithiasis, strictures including those due to primary sclerosing cholangitis, intrahepatic biliary stones,23 AIDS cholangiopathy,24,25 pancreatitis, neoplasia of the biliary tree and Caroli’s disease. If cholangitis persists or recurs over a period of years, secondary biliary cirrhosis may develop. The histological features are then as described above in the section on bile duct obstruction. Septicaemia uncommonly is associated with a particular form of histological cholangitis principally affecting the canals of Hering.26 Affected ductules are dilated and filled with inspissated bile. Neutrophils accumulate around and sometimes within them. Larger ducts may be affected, as may the periportal parenchyma in which bile is seen in dilated bile canaliculi. These changes are easily confused with those of large bile-duct obstruction, but in obstruction the inspissated bile in the canals of Hering is not a feature unless there is concomitant sepsis. Sepsis more often gives rise to widespread canalicular cholestasis; the ductular cholestasis pattern (see Fig. 15.11) is seen in the minority of septic patients.26 In the toxic shock syndrome the appearances of the small bile ducts can closely mimic ascending bacterial cholangitis.27
Primary sclerosing cholangitis (PSC) is characterised by inflammation, strictures and saccular dilatations in the biliary tree. Typically found in adults with ulcerative colitis, it is also seen in neonates and children28 and in the absence of inflammatory bowel disease. In a few cases the latter is Crohn’s disease rather than ulcerative colitis.29 Any part of the biliary tree may be affected, and involvement of the gallbladder30 and pancreas31 has been reported. The gallbladder shows intramural lymphoplasmacytic infiltrates and lymphoid aggregates.32 Patients do not necessarily have symptoms referable to the liver or abnormal liver function tests.33 The disease may recur after liver transplantation.34 Lesions similar to those of primary sclerosing cholangitis have been found in patients given arterial infusion of the anti-cancer drug fluorodeoxyuridine and other35 chemotherapeutic agents.36 Obliteration or narrowing of hepatic arteries and portal-vein branches suggests that, in drug-related cases at least, the bile-duct damage may have an ischaemic origin.37 Systemic vasculitis, liver transplantation-related hepatic artery thrombosis or chronic rejection vasculopathy and, rarely, septic shock38 are other causes of ischaemic bile-duct injury39 (ischaemic cholangiopathy).
Final diagnosis of primary sclerosing cholangitis normally rests on cholangiographic demonstration of the characteristic beading of bile ducts, but similar histological features, as described below, may be found in patients with normal cholangiograms. This can be explained on the basis of involvement of the smallest ducts, too small to be seen radiographically.40 This small-duct primary sclerosing cholangitis corresponds approximately to the now obsolete label of ‘pericholangitis’, when applied to patients without cholangiographic abnormalities. Large- and small-duct forms of the disease frequently coexist.
The features seen on liver biopsy depend in part on the location of strictures in relation to the biopsy site. If the biopsy is taken from a part of the liver unaffected by the primary disease but proximal to a stricture, then the changes, if any, will simply be those of bile-duct obstruction or cholangitis. The presence of chronic inflammation may lead to confusion with chronic hepatitis. If, on the other hand, the biopsy site is affected by the primary disease, there may be one or more features suggesting the diagnosis. These include periduct oedema and concentric fibrosis (Fig. 5.14), ductular proliferation, portal inflammation and atrophy or disappearance of the small ducts (Fig. 5.15). Loss of ducts is the commonest finding in the smallest portal tracts, while periduct fibrosis is typical of medium-sized tracts.41 Major bile ducts, as seen for example in explanted livers at transplantation, may be inflamed, ulcerated or dilated. They may also rupture, producing a perihilar xanthogranulomatous cholangitis.42
Figure 5.14 Primary sclerosing cholangitis.
Figure 5.15 Primary sclerosing cholangitis.
The inflamed portal tract lacks a bile duct. An aggregate of lymphocytes to the left of a small hepatic arteriole at the centre of the field is likely the former site of the duct. Inflammation extends into the adjacent parenchyma and there is interface hepatitis to the right of the portal tract. (Wedge biopsy, H&E.)
Loss of interlobular bile ducts from the smallest portal tracts can be assessed only in biopsy samples of adequate size, that is to say, containing several portal tracts. While interlobular ducts are not necessarily seen in all tracts because of the plane of section, arteries provide a useful guide; from 70–80%43 to 92%44 of arteries are normally accompanied by a duct lying near the centre of a portal tract. If there is doubt, for instance because ducts are difficult to identify in an inflammatory infiltrate, immunostaining of duct-associated cytokeratins is helpful.45 Suitable antibodies include AE-1 (Signet) and other antibodies against cytokeratins 7 and 19. In the presence of a ductular reaction, identification and counting of interlobular bile ducts are sometimes difficult.
The concentric fibrosis around medium-sized ducts is not entirely diagnostic, since it is occasionally found in other forms of biliary disease such as hepatolithiasis.46 It is, however, a very helpful finding. The lamellar pattern of the fibrosis gives an ‘onion-skin’ appearance. The cuff of connective tissue around the duct may be oedematous and pale-staining or sclerotic, depending on the stage of the process. Inflammatory cells are seen in small numbers lying between the layers of collagen. The duct epithelium may show various degrees of atrophy, and sometimes disappears entirely, leaving a characteristic rounded fibro-obliterative scar47 (Fig. 5.16). Staining with diastase–PAS often reveals irregular or regular thickening of the basement membrane material around both scarred and unscarred ducts.48 In long-standing or severe cases, portal fibrosis gradually increases, fibrous septa form and secondary biliary cirrhosis may develop. In some patients, on the other hand, the lesions remain mild and clinically insignificant for many29 years.33 Portal tract fibrogenesis in sclerosing cholangitis and in primary biliary cirrhosis is in part attributed to an increased number of intrahepatic mast cells compared with other chronic liver diseases.49 In fact, systemic mastocytosis has been associated with cholestasis50 and a case of PSC.51
Figure 5.16 Primary sclerosing cholangitis.