Vol. 15, No. 1 January/February 1991

0145-6008/9 1/150 1-0045$3.00/0 ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH

Alcoholic Liver Disease: Pathologic, Pathogenetic and Clinical Aspects Kamal G. Ishak, Hyman J. Zimmerman, and Mukunda B. Ray

Alcoholic liver disease includes steatosis, alcoholic hepatitis and cirrhosis. Other liver diseases of genetic origin, but with a curious association with alcohol intake, are hemochromatosisand porphyria cutanea tarda. The attribution of chronic hepatitis to alcohol intake remains speculative, and the association may reflect hepatitis C infection. Hepatic injury attributed to alcohol includes the changes reported in the fetal alcohol syndrome. Steatosis, the characteristic consequence of excess alcohol intake, is usually macrovesicular and rarely microvesicular. Acute intrahepatic cholestasis, which in rare instances accompanies steatosis, must be distinguished from other causes of intrahepatic cholestasis (e.g., drug-induced) and from mechanical obstruction of the intrahepaticbile ducts (e.g., pancreatitis, choledocholithiasis) before being accepted. Alcoholic hepatitis (steatonecrosis) is characterized by a constellation of lesions: steatosis, Mallory bodies (with or without a neutrophilic inflammatory response), megamitochondria, occlusive lesions of terminal hepatic venules, and a lattice-like pattern of pericellular fibrosis. All these lesions mainly affect zone 3 of the hepatic acinus. Other changes, observed at the ultrastructural level, are of importance in progression of the disease. They include widespread cytoplasmic shedding, and capillarization and defenestration of sinusoids. Progressive fibrosis complicating alcoholic hepatitis eventually leads to cirrhosis that is typically micronodular but can evolve to a mixed or macronodular pattern. Hepatocellular carcinoma occurs in 5 to 15% of patients with alcoholic liver disease. The clinical syndrome of alcoholic liver disease is the result of three factors-parechymal insufficiency, portal hypertension and the clinical consequences of extrahepatic damage produced by alcohol. At the several phases of the life history of alcoholic liver disease, the individual factors play a different role. The clinical manifestations of alcoholic steatosis are mainly extrahepatic in origin. Those of alcoholic hepatitis reflect mainly parenchymal insufficiency and those of cirrhosis are mainly those of portal hypertension. Alcoholic liver injury appears to be generated by the effects of ethanol metabolism and the toxic effects of acetaldehyde, perhaps the immune responses to alcohol- or acetaldehyde-altered proteins, and questionably enhanced by viral hepatitis. Alcoholic hepatitis may be mimicked histologically, and to a varying degree clinically, by a number of conditions (obesity, diabetes, several drug-induced injuries, jejunoileal bypass, and related “shortcircuiting” of the bowel). Perhaps the most important facet of the hepatotoxicity of alcohol is its enhancement of the effects of a number of other hepatotoxic agents, among which acetaminophen

is the prime example.

lcoholic liver disease (ALD) is expressed mainly in A three major clinico-pathologic settings; alcoholic steatosis (alcoholic fatty liver), alcoholic hepatitis, and alcoholic fibrosis and cirrhosis. Other manifestations include alcohol-induced cholestasis, chronic active hepatitis, fetal alcohol syndrome and hepatocellular carcinoma. In addition, there is a curious association with iron-overload disease. Up to 40% of those with hemochromatosis are alcoholic, and alcohol appears to contribute to the development of porphyria cutanea tarda in about 70% of patients with that condition. It is also worth remembering that there is a complex interrelationship between alcohol and other hepatotoxins;’ and alcohol may enhance the toxic effects of a number of these. Liver injury, for example, can occur in alcoholics on therapeutic doses of acetaminopheq2 the enhancement may be related to glutathione deficiency associated with alcoholism and to induction of cytochrome P450 by ethan01.~Acute viral hepatitis may be superimposed on ALD4-6distorting the latter; and there is a high incidence of hepatitis B and C markers in ALD.7,8 This review covers the pathologic effects of acute and chronic injury to the liver induced by alcohol, with clinical and pathogenetic correlations. HISTOPATHOLOGY

A Ecoholic Steatosis This occurs in the majority of heavydrinkers, but is considered a reversible lesion and not necessarily a precursor to alcoholic hepatitis.’-’ Alcoholic steatosis is either macrovesicular or microvesicular, but admixtures of the two types may occur. The hepatic lipids accumulating in patients with ALD (and also in patients with morbid obesity) are not only triglycerides but also fatty From the Department of Hepatic Pathology (K.G.I.) and American acids, monoglycerides and diglycerides.l 2 Registry of Pathology (H.J.Z.), Armed Forces Institute of Pathology, Macrovesicular steatosis is typified by the presence of Washington, DC: and University of Cincinnati Medical Center, Cincigone medium-sized to large fat droplet (vacuole in HE nati, Ohio (M.B.R.). section) per hepatocyte, with lateral displacement of the Receivedfor publication June 28, 1990: accepted July 17, 1990. *The opinions or assertions contained herein are the private views of nucleus (“signet-ring” hepatocyte) (Fig. 1). The fatty the authors and are not to be construed as official or as rejecting the change is often pan-acinar, but may be more prominent views ofthe Department ofthe Army or the Department of Defense. in zone 3. Lipogranulomas, attributed to rupture of fat Reprint requests: Kamal G. Ishak, Ph.D., Chairman, Department of Hepatic Pathology, Armed Forces Institute of Pathology, Washington, “cysts,” are rare in our experience, but are purported to occur in 30% to 50% of cases of ALD.6 NecroinflammaDC 20306-6000. Copyright 01991 by The Research Society on Alcoholism. tory changes, cholestasis, and fibrosis are absent or very

AlcoholClm ExpRes, Vol 15, No 1, 1991: pp45-66

’

45

46

mild. Megamitochondria, which appear as eosinophilic globules of varied size that are PAS-negative, are typically present in acinar zone 3; more elongated needle-shaped mitochondria may be seen in zone 1.13-15 Giant mitochondria also occur in other types of ALD, e.g., alcoholic hepatitis, as will be noted later. They are nonspecific and found in a variety of liver diseases. Microvesicular steatosis in ALD, also dubbed “alcoholic foamy degeneration” by Uchida et a1.6 and other~,”~~’ is characterized by a striking increase in the size of liver cells, which are filled with small fat droplets (Fig. 2). The cells have centrally placed nuclei; bile may accumulate in their cytoplasm and in canaliculi. Focal necrosis may be found in the vicinity of terminal hepatic venules. A decrease or absence of the activity of several enzymes was demonstrated by Uchida et al. l 6 Microvesicular steatosis is a rare lesion and is reported in 0.8% and 2.3% in two series of ALD from France” and Spain,” respectively. The morphologic differential diagnosis of steatosis in alcoholic liver disease is listed in Table l . Focal fat infiltration is a well known entity that is recognized by various imaging technique^.'^,^^ It occurs most frequently in ALD, but is also seen in other conditions leading to hepatic steatosis. Several patterns of infiltration have been described” and the changes are rapidly reversible after abstinence from alcohol. Histologically, the acinar architecture is preserved. The steatosis is macrovesicular and involves varied numbers of multiple contiguous acini.

Alcoholic Hepat itis A synonym for this condition is alcoholic steatonecrosis, a term introduced in 1955.*’ In its advanced form alcoholic hepatitis has been called “sclerosing hyaline necrosis.” 22 It is a chronic lesion that predominantly affects acinar zone 3. The reason for the predominance of the liver injury in zone 3 may be the increased production of ethanol-derived toxic metabolites that in turn is due to elevated levels of P450IIEI in liver cells in that zone.23 Several comprehensive reviews of alcoholic hepatitis, as well as other histopathologic and pathogenetic aspects of ALD are recommended for further reading.6,24-29 Conditions morphologically mimicking ALD are listed in Table 1. Microscopically, alcoholic hepatitis is characterized by a constellation of lesions that vary in degree and extent from patient to patient. In addition to fatty metamorphosis (usually macrovesicular but sometimes “mixed”), there is ballooning degeneration of liver cells that is most prominent in acinar zone 3, with Mallory body formation. The Mallory bodies (alcoholic hyaline) are eosinophilic and may be short and irregular, or long and rope-like; when long, they can form a complete or open ring around the nucleus (Figs. 3 and 4). The cytoplasm around the Mallory bodies is typically empty or rarified (Fig. 3 ) . It iS pertinent at this Juncture to briefly comment on

ISHAK ET AL.

the ultrastructure and pathogenesis of Mallory bodies. Three types have been described ~ltrastructurally.~~ Type I consists of bundles of filaments (5-20 nm thick) in parallel arrays; type I1 is composed of randomly oriented fibrils (Fig. 5);type 111 consists of a granular or amorphous, markedly electron-dense substance containing only scattered fibrils. Simply stated, Mallory bodies are thought to result from a derangement of the intermediate filament component of the cytoskeleton of liver cells. Several theories regarding their pathogenesis have been advanced: microtubular failure, preneoplasia with a structural phenotypic alteration, vitamin A deficiency, and membrane injury by free fatty acid^.^'-^^ Monoclonal antibodies to cytokeratin that can stain Mallory bodies are commercially available (Fig. 6). Using such antibodies Ray3’ detected Mallory bodies by the PAP technique in 71% of cases but only in 40% by the hematoxylin and eosin stain. Monoclonal antibodies directed against Mallory bodies

Fig. 1. Macrovesicular steatosis, as well as fibrosis in Zone 3, in patient with alcoholicliver disease (HE x160).

Fig. 2. Microvesicular steatosis. Many small fat vacuoles are present in each hepatocyte. Note centrally located nuclei (HE x630).

ALCOHOLIC LIVER DISEASE

41

Table 1. Morphologic Differential Diagnosis of Alcoholic Liver Injury Pattern of alcoholic injury Macrovesicular steatosis Microvesicular steatosis Acute intrahepatic cholestasis

4. Alcoholic hepatitis

5. Alcoholic chronic active hepatitis (CAH) 6. Fibrosis (zone 3)

7. Fibrosis (zone 3) with venocclusive lesions

8. Cirrhosis 9. Cirrhosis with iron overload

Examples of diseases with similar morphology Obesity, diabetes mellitus, corticosteroid therapy Tetracycline or salicylate toxicity, acute fatty liver of pregnancy Large duct obstruction (e.g., acute pancreatitis, choledocholithiasis), drug-induced cholestasis. Zieve’s syndrome Pseudoalcoholic hepatitis (nonalcoholic steatohepatitis) associated with diabetes mellitus and/or obesity, jejuno-ileal bypass or drug injury (e.g., perhexilene maleate, amiodarone, nifedipine, diltiazem. 4,4-diethylaminoethoxyhexestrol, diethylstilbestrol) Non-A, non-B CAH, autoimmune CAH, drug-induced CAH HypervitaminosisA, methotrexate therapy, diabetes mellitus Venocclusive disease secondary to pyrrolizidine alkaloids, therapeutic drugs (e.g., dacarbazine), or radiation injury Cirrhosis complicating conditions causing pseudoalcoholichepatitis (non-alcoholic steatonecrosis) Genetic hemochromatosis

Fig. 4. Mallory body forms an incomplete ring around the nucleus of a liver cell (HE ~ 1 0 0 0 ) .

Fig. 3. Ballooned liver cells have an empty cytoplasm except for the presence of Mallory bodies (HE x750).

appear to be even more sensitive. In one study they detected Mallory body material in 77% of liver biopsies from patients with ALD, as compared with 23% of liver biopsies stained by conventional methods.38Another antibody, MM 120-1, that is highly specific for murine and human Mallory bodies, was recently reported by Zaloukal et it recognizes a high molecular weight nonkeratin component of Mallorv bodies. Death of the cell harboring the Mallory body is associated with nuclear and cell membrane lysis and excites a neutrophilic inflammatory cell response (Fig. 7); sometimes a ring of neutrophils surrounds the Mallory body (“satellitosis”).Lymphocytic sequestration by the liver in

Fig. 5. Mallory body consists of non-membrane bound, randomly oriented fibrils in the cytoplasm of a hepatocyte (Electron micrograph x*4,000),

alcoholic hepatitis has been observed ultrastructurally in two patient^.^' The interaction between neutrophils and hepatocellular injury in ALD appears to be quite complex. Hepatocytes incubated with ethanol in vitro produce a chemotactic

ISHAK ET AL.

48

Fig. 6. Mallory bodles are tmmunostatned with monoclonalantibodies AE1IAE3 to cytokeratin (x250).

Fig. 8. Two liver cells each contain one sphertcal megamitochondrion(arrows) (HE XIOOO).

Fig. 7. Intra-acinar inflammatory response in alcoholic hepatitis (~250).

factor for Kupffer cells and n e ~ t r o p h i l s . ~ Neutrophils '~~~ migrate into liver cells containing Mallory bodies; their degranulation and collapse is thought to be one of the factors contributing to the hepatocellular damage in ALD.43Free radical generation by neutrophils is another possible mechanism of cell injury in ALD proposed by one group of investigator^,^^ but is not supported by the recent studies of Hayes et al.45 Megamitochondria may be present in alcoholic hepatitis. In one series of 208 patients with ALD, megamitochondria were found in 28% of patient^.'^ Most often they are round and eosinophilic and are usually located in zone Fig. 9. Megamitochondriahave irregular shapes and contain paracrystallineand harboring a granular inclusions. They also show loss of cristae (Electron microqraph ~24,000). . 3'3 - .. (Fig. '-)* They - _ _ _can be Seen i n the Same Mallory body. Ultrastructurally, several types have been de~cribed:'~ Type I is spherical, with a paucity of cristae ALD with megamitochondria had a more benign course (Fig. 9); type I1 is elongated and has long crystalline than those inclusions; and type 111 is small and often bizarre in shape, Megamitochondria must be distinguished from alphawith multiple crystalline inclusions. In one study, cases of 1-antitrypsin (AAT) globules; the former are PAS-nega-

ALCOHOLIC LIVER DISEASE

tive. Pariente et al.47reported PAS-positive AAT globules in seven patients with alcoholic cirrhosis who had a normal M phenotype. However, as pointed out by Roberts et al.48 some patients with an apparently normal AAT phenotype may have a defective M variant, such as M malton. After abstinence from alcohol, fatty metamorphosis resolves in 3 to 4 weeks while Mallory bodies are said to disappear in 6 to 12 weeks in humans.23In the griseofulvin-treated mouse, Mallory bodies have been found to be very durable structures; they have been found for up to 6 months after griseofulvin ~ i t h d r a w a l . ~ ~ ” In addition to Mallory bodies, there are focal necroses and scattered acidophilic/apoptotic bodies in alcoholic hepatitis (Fig. 10). In one study acidophilic bodies were observed in 30% of biopsies from patients with alcoholic hepatitis.49They have also been found in the livers of rats given ethanol for 5 weeks.50 Cholestasis, if present, is usually minimal but can be prominent when liver failure or sepsis occurs. Other features that are quite variable are periportal cholangiolar pr~liferation,~’ hemosiderosis (hepatocellularand/or reticuloendothelial), and oncocytic (oxyphil), or ground-glass (“induced”) liver cells. About one third of chronic alcoholics have an increased hepatic

Fig. 10. Sinusoidal acidophilic body is electron dense and has an irregular outline. Organelles are not recognizable, Several lipid droplets are present in adjacent hepatocyte (top) (Electron micrograph ~6,800).

49

iron c~ncentration.~~ The multiple factors that can contribute to iron overload in the alcoholic are addressed in an editorial by A i ~ e n Kupffer . ~ ~ cells may show mild to moderate hypertrophy, with or without hemosiderin accumulation. There is general agreement that Kupffer cells, both in ALD54,56and ethanol-fed show impaired function, as measured by decreased phagocytic activity or diminished lysozyme content. Continued activity of the lesions of alcoholic hepatitis leads to progressive pericellular fibrosis in acinar zone 3 (Figs. 11 and 12), with a lattice-like or “chicken-wire” appearance in sections stained with connective tissue stains,23periportal fibrosis and occlusive lesions of terminal hepatic venules (THV). Almost all patients with alcoholic steatonecrosis or cirrhosis have phlebosclerosis of THV, a lesion originally described by Edmondson et a1.22 Other venular lesions include a lymphocytic phlebitis5’ and venocclusive lesion^'^-^^ (Fig. 13). While the most striking venular lesions affect the THV, portal venous occlusive lesions have also been observed in alcoholic hepatitis and c i r r h ~ s i s . ~The ~ ~ ~etiopathogenesis ’ of the vascular lesions and their effect on evolution of the ALD require further study. As already noted alcoholic fibrosis is a recognized sequel to alcoholic steatonecrosis, but it can occur in its absence. Perivenular fibrosis in ALD is reported to have predictive value for subsequent progression to cirrhosis.62The fibrosis is associated with fatty metamorphosis, and can develop in the absence of widespread inflammation and necrosis.63 The occurrence of liver fibrosis in the alcoholic can be detected by Fab radioimmunoassay of serum procollagen I11 pep tide^.^^ Nasrallah et al.65 have emphasized that perivenular fibrosis is almost always “pericellular.” In fact, Nasrallah et al.65and Caulet et a1.66believe that pericellular fibrosis is of greater prognostic significance in ALD than is perivenular fibrosis; the latter investigators used morphometric as well as histologic grading methods. Increased

Fig. 11. Extensive fibrosis in zone 3 with disruption of the liver cell plates. 50). Lumen of terminal hepatic venule is indicated by arrow head (HE XI

50

Fig. 12. Marked fibrosis in space of Disse. Thick bundles of collagen appear to be randomly oriented (Electron micrograph ~10,000).

Fig. 13. Terminal hepatic venule is partially occluded by intimal thickening. Note fibrous septa extending outward from vein (HE x80).

ISHAK ET AL.

accumulation of fibronectin may act as a chemotactic factor for neocollagen formation.67The fibrosis appears to be related to proliferation of fibroblasts and myofibroblasts, both of which may represent transformed perisinusoidal lipocytes or It0 a finding also noted in alcoholic liver injury in the baboon.70Additionally, highly significant lesions have been detected by both transmission and scanning electron microscopy in acinar zone 3 by Horn et al.71372 They include deposition of basal laminalike material and defenestration of the wall of sinusoids. The “capillarization” (fibrosis and basement membrane formation) of sinusoids is associated with accumulation of Factor VIII-related antigen and fibr~nectin.’~ At this juncture note should be made of the deposits of IgA that are found in ALD; they are present in a continuous pattern in a perisinusoidal l o ~ a l i z a t i o n . Several ~~-~~ groups of investigator^^^,^^ believe that they represent a distinct effect of alcohol on the liver related to the role of this organ in IgA metabolism. Others have suggested that the IgA deposition is not specific for alcoholic liver disease,76,77 but may reflect the impaired metabolism of the damaged liver. Another recently observed phenomenon is the induction of polypeptides termed “heat shock proteins” (HSP) in ALD. One of these, an 8.5-kd polypeptide known as ubiquitin, has been identified immunohistochemically in Mallory bodies in human ALD78 and in griseofulvininduced Mallory bodies in mice.79More recently, another 70-kd HSP was demonstrated immunohistochemically in hepatocytes, including those not showing Mallory bodies or steatosis, in patients with ALD.79aA possible role of this HSP in the cytoskeletal derangements of ALD has been suggested by Omar et al.79a The conclusion to be drawn from the published data briefly mentioned in this section is that significant fibrosis (perivenular and perkellular) can occur in alcoholics in the absence of the necroinflammatory changes collectively referred to as alcoholic hepatitis. In Japan, less than 10% of alcoholics appear to develop alcoholic hepatitis, the remainder having hepatic fibrosis, steatosis with fibrosis, chronic active hepatitis and cirrhosis, (with or without superimposed carcinoma) without having gone through a stage of alcoholic hepatitis.80,81A recent review of hepatic fibrosis caused by alcohol is recommended for further reading.82The morphologic differential diagnosis of fibrosis in ALD is listed in Table 1.

Alcoholic Cirrhosis With progression of alcoholic hepatitis, fibrous septa begin to link the chicken-wire fibrosis in zone 3 to extensions of the periportal fibrosis (Fig. 13), eventually leading to complete encirclement of islets of hepatic parenchyma. The cirrhosis that develops is usually micronodular (Fig. 14). A macronodular pattern can evolve after alcohol withdrawa1’83-86When hepatocellular carcinoma it does so in the setting of macronodular ~ i r r h o s i s . ~ ~ , ~ ~

ALCOHOLIC LIVER DISEASE

51

to hepatitis C may be the cause of chronic hepatitis in alcoholics, as determined in their patients by serologic testing for anti-HCV; in that study 35% of the alcoholics were anti-HCV positive. In another more recent series from Italy, anti-HCV was detected in 52% of patients with a history of heavy alcohol intake."' However, a report of two alcoholic patients with chronic active hepatitis, whose biochemical findings and necroinflammatory disease activity improved after abstinence from alcohol, lends credence to a direct relationship between alcohol and chronic hepatitis, at least in some cases."* One possible mechanism for autoimmune reactions and the development of chronic active hepatitis is the reduced phagocytic activity of macrophages that leads to shedding of hepatocellular material into the 3. Hepatocellular Carcinoma. This occurs in 5 % to 15% of patients with ALD. The subject has been extensively reviewed by several authorities and is not discussed further Fig. 14. Micronodular cirrhosis of alcoholic eiology. Note absence of necroinby us.27, 114-123 flammatory activity (HE x60). 4. Fetal Alcohol Syndrome. A pattern of altered growth There are numerous conditions that may mimic alcoand morphogenesis can occur in the offspring of alcoholic holic steatonecrosis (Table 1). The differential diagnosis is mothers. The disorder, termed fetal alcohol syndrome, is discussed in the previously cited review^.^^-*^-^^ The two characterized by growth deficiency, behavioral abnormalmost clinically relevant entities are nonalcoholic steatoities, microcephaly, craniofacial and limb abnormalities, necrosis related to obesity and/or diabetes m e l l i t ~ s ' ~ - ~ ~ and cardiac septa1 defect^.'^^-'^^ Growth deficiency and and amiodarone h e p a t o t ~ x i c i t y . ~Other ~ - ~ ~drugs leading dysmorphism have been reported to continue during to pseudoalcoholic hepatitis that should be mentioned childhood in one long-term study.127 perhexilene include 4,4-diethylaminoethoxyhexe~trol,~~ Hepatic dysfunction and histopathologic abnormalities maleate,98d i l t i a ~ e m nifedipine,'" ,~~ and diethylstilbeshave been found in five patients with the fetal alcohol trol. ~ y n d r o m e . ' ~ ' -One ' ~ ~ patient had congenital hepatic fibrosis,128a condition that is of autosomal recessive inheritance. That patient and another had sclerosis of terminal Other Conditions I . Acute Alcoholic Cholestasis, a term proposed for a hepatic venules.12' Four of the patients reported by Habsyndrome following excessive drinking, is characterized bick et a1.I2' showed mild to marked steatosis. Bile duchistopathologically by severe cholestasis in zone 3, in the tular proliferation was noted in two of the iivers.129,130 absence of alcoholic hepatitis. Io3 Glover et a1.,Io2who Ultrastructural study of one liver revealed collagen and reported three patients with this syndrome, emphasized basement membrane deposition in spaces of Disse, tothe heterogeneity of cholestasis in the alcoholic patient, gether with the presence of perisinusoidal cells and myoHepatoblastoma has been reported in a 27including its occurrence in alcoholic steatosis, alcoholic fibr0b1asts.l~~ month-old child who had the fetal alcohol syndrorne.I3' hepatitis, and Zieve's syndrome. Afshani et al. Io4 have However, that child had also received azathioprine for 20 attributed severe cholestasis in ALD to an intrahepatic months prior to death. While the aforementioned in"microscopic cholangitis." Other causes, such as acute or stances of possible liver cell damage are intriguing, it chronic pancreatitis, with common bile duct s t e n o ~ i s ' ~ ~ should be borne in mind that similar changes have not choledocholithiasis, or superimposed viral or drug-in100 patients in two series of the been identified in over duced cholestasis, also must be considered in differential fetal alcohol s y n d r ~ r n e . 'Many ~ ~ ' ~more ~ ~ cases need to be diagnosis of cholestasis in the alcoholic patient (Table 1). studied before the hepatic biochemical and histopatho2. Alcoholic Chronic Active Hepatitis. There are several logic abnormalities of this syndrome can be clarified. reports of alcohol-related chronic active hepatitis in the Western literature. 106-110 As already noted the data are quite different in Japan. In three series of ALD in Japan reviewed by Ohnishi and Okuda,"' the incidence of CLINICAL SYNDROMES chronic active hepatitis was lo%, 16%, and 27%, respectively. Obviously, a non-A, non-B (hepatitis C) viral etiolA mountain of literature deals with the clinical maniogy can only be ruled out as serologic tests become com- festations of alcoholic liver disease; almost all body sysmercially available. In a recent study from Italy, Brillanti tems may seem i n ~ o l v e d . ~ ~ Overt , ' ~ ~ -hepatic ~ ~ ' disease and collaborators109have suggested that sporadic exposure (hepatomegaly,jaundice, ascites, and portal hypertension)

ISHAK ET AL.

52

may be accompanied by digestive, endocrine, hematologic, muscular, and other extrahepatic manifestations (Table 2). These constitute a constellation of clinical features, some due to parenchymal injury, others to distorted portal circulation, some to both factors acting concurrently, and many to extrahepatic organic sequelae of alcoholism rather than the liver disease (Fig. l 5).134-137 Parenchymal injury with failure of hepatic function appears to account for jaundice, hypoalbuminemia, depressed plasma coagulation factors, and some of the endocrine stigmata, and perhaps contributes to systemic manifestations, such as weight loss and fever. Distorted portal circulation (and resultant portal hypertension) leads to esophageal varices and, by producing splenomegaly and hypersplenism, may lead to anemia, leukopenia, and thrombocytopenia. Parenchymal injury and portal obstruction, together, are responsible for hepatic encephalopathy and ascites. Extrahepatic organic sequelae of alcoholism are often responsible for digestive and other complaints that are attributed to ALD.". 134Nausea, vomiting, and abdominal pain may reflect gastritis or pancreatitis. Folate deficiency, that is frequently associated with alcoholism, may contribute to anemia, leukopenia, and thrombocytopenia. Alcoholism also can produce myopathy, neuropathy, and cardiomyopathy, and neuropsychiatric abnormalities that may merge with effects of the hepatic d i ~ e a s e . ' ~ ~ - ' ~ ' Table 2. Clinical Features of Patients with Alcoholic Cirrhosis Clinical manifestations

System or organ involved ~

Hepatic Digestive Endocrine Hematologic Renal and electrolytes Circulatory Neurologic

Systemic PATHOLOGY:

Analysis of the contribution of the separate vectors, (parenchymal hepatic injury, portal hypertension, and extrahepatic organic sequelae of alcoholism) provides a guide to the clinical manifestations of alcoholic cirrhosis at any point in its life history (Table 3). The initial lesion of ALD is the fatty liver. In some alcoholics, this is followed by cirrhosis, which progresses through early stages to ultimate development of advanced disease. Along the course of the disease, alcoholic hepatitis develops as a complication of the fatty liver or of evolving cirrhosis.

Hepatic Steatosis Fatty liver is found in 70% to 100% of patients taking 135, 14' A lesion of only excessive amounts of alcoh01.~~ minor clinical importance per se, its chief importance is as the first histologic evidence of the adverse effects of alcohol on the liver, recognizable by light rnicro~copy.'~~ There are few clinical manifestations clearly attributable 142 They include hepatomegaly with to the fatty or without endocrine stigmata (spider nevi, gynecomastia, testicular atrophy). There is little or no jaundice, ascites, significant portal hypertension, or liver failure. The cholestatic syndrome that has been attributed to severe '03, 143is a rare phenomenon. Pancreatitis, leadsteatosis102, ing to obstruction of the biliary tree, lo4, 144may be responsible for some of the instances of jaundice in the alcoholic patient attributed to the fatty liver. Indeed, clinical and biochemical manifestations of alcoholic steatosis reflect mainly extrahepatic sequelae of alcoholism. 134,145 Biochemical evidence of hepatic injury is usually light.^^,'^,'^ Mild hyperbilirubinemia, usually subicteric, is found in about 20% of patients. Higher values are rare. Serum enzyme levels are only mildly elevated. The AST value is rarely more than twice normal and the ALT level is even 10wer.'~~''~' Serum globulin levels are normal or mildly elevated, occasionally low.'41 Delirium tremens or alcoholic myopathy may lead to marked elevations of aminotransferases, lactic dehydrogenase, and creatine kinase. 146 Hypoprothrombinemia is an unusual manifestation; when present, it suggests that the hepatic lesion includes more than simple steatosis. The prognosis of the patients with a fatty liver is good.'0,'45,14' It depends on the severity of other diseases 3'41

3''

Jaundice, ascites. varices, fetor hepaticus Anorexia, vomiting, abdominal pain Gynecomastia spider nevi, testicular atrophy Coagulopathy, anemia, leukopenia, thrombocytopenia Azotemia, alkalosis-acidosis, hypokalemia, hyponatremia Cyanosis, clubbing, hyperkinetic circulation Encephalopathy-coma (and various other neuropsychiatric sequelae of alcoholism) Wasting, fever PARENCHYMAL INJURY

ARCHITECTURAL DISTORTION (fibrosis and pseudolobule formation)

-----

Table 3. Relative Importance* of Facets of Alcoholic Liver Disease at Each Stage Stage

CLINICAL:

PORTAL HYPERTENSION

PARENCHYMAL FAILURE,

/

\1

d

SYNDROME OF CIRRHOSIS

f ASSOCIATED DISEASES (e.g. Alcoholism) Fig. 15. Components of syndrome of cirrhosis. From Zimmerman (133).

Steatosis Steatonecrosis (alcoholic hepatitis) Cirrhosis

Parenchymal failure

Portal hypertension

Extrahepatict manifestations

+

4+

1-2+

4+ 1-4+

1+

4+

1-4+

f

* 0, No importance; A , not prominent, relatively little importance; 1+, usually present, plays minor role in syndrome; 4+, most important aspect of syndrome. t Anorexia, nausea, vomiting, abdominal pain due to alcoholic gastritis and pancreatitis, hematologic abnormalities due to marrow suppression by ethanol or folate deficiency; endocrine manifestations due to effect of ALD on hormonal metabolismand to alcoholic injury to gonads; myopathic and neuropathic and other extrahepatic organic manifestationsof alcoholism.

ALCOHOLIC LIVER DISEASE

such as alcoholic pancreatitis, gastritis, delirium tremens, or pneumonia. Instances of sudden death, perhaps as a manifestation of fat embolism, however, occur.” lo- 145*148 Alcoholic microvesicular steatosis, (“alcoholic foamy liver”)an uncommon lesion described in an earlier section presents with jaundice and few other complaints. I5-l8 The prognosis is generally good.

Alcoholic Hepatitis This stage of ALD has been estimated to occur in 17% to 42% of alcoholics.’49, The classical clinical syndrome consists of jaundice, varying degrees of hepatic failure, abdominal distress, fever, and leucocytosis, although patients with the histological features of alcoholic hepatitis may be asymptomatic and a n i ~ t e r i c . ’IS2~ ~The , syndrome appears in patients who have been consuming excess amounts of alcohol for periods of 5 years or more (rarely as short as 1 year), and generally follows periods of particularly heavy intake of alcoh01.~’ Alcoholic hepatitis may be regarded as a complication of the fatty liver or cirrhosis.21Accordingly, it may be found in a liver that contains no evidence of cirrhosis or in one with established c i r r h o ~ i s . * ~In~ the ~ ~ ~former ~’ instance, the clinical manifestations would be those of parenchymal injury without portal hypertension. In the latter instance, the clinical picture is a composite of features that reflect both parenchymal injury and portal hypertension. The dominant clinical features of the syndrome of alcoholic hepatitis reflect the parenchymal insufficiency. In general, the severity of the degeneration and necrosis determines the depth ofjaundice, the severity of coagulopathy and the degree of hypoalbuminemia. (Coincidental pancreatitis causing obstruction of the biliary tree, however,104,144 may lead to more intense jaundice than the degree of parenchymal injury would be expected to cause.) Parenchymal insufficiency also contributes to the development of ascites and coma. Endocrine stigmata reflect the parenchymal injury, 134 although testicular atrophy may also be the result of a toxic effect of ethanol on the testes.154Splenomegaly, an abdominal collateral venous pattern, and esophageal varices, as well as ascites and hepatic encephalopathy, are found in varying proportions of patients with alcoholic hepatitis, depending largely on the degree of cirrhosis. Nevertheless, portal hypertension may develop in alcoholic hepatitis even without cirrhosis depending, presumably, on the degree of perivenous sclerosis22as well as the effect of sinusoidal blockade or 155-’s7 or sinusoidal inflexibility due to ~apillarization’~~ compression by enlarged hepatocytes.15*Encephalopathy may be the result of parenchymal injury, portal venous shunting, or both, and may be spontaneous or provoked by gastrointestinal hemorrhage.*I Laboratory data usually show anemia and, frequently, leucocytosis. Values for AST are elevated in more than 90% of patients, but in almost all of them the level is less

53

than eight times the normal. Values for ALT are even lower than those for AST in almost all patients with alcoholic hepatitis.49* 146 Indeed, the ratio of AST:ALT values range from 8: 1 to 2: 1, a phenomenon useful in the clinical recognition of the syndrome. Hypoalbuminemia is seen in about 70% of these patients, and hyperglobulinemia of modest degree in more than half of them.49 The prognosis of alcoholic hepatitis is better than that accorded it in early views of the syndrome. Older views,’59 which were influenced by data from necropsy studies, gave the condition a grim prognosis. More recent data, based on biopsy studies,49,151-153,160-164 have yielded a relatively favorable view of the immediate prognosis, since prothrombin values that permit liver biopsy are usually (90% to 95% of patients) associated with recovery from the acute episode. Adverse prognostic markers include, in addition to abnormal coagulability, any degree of azotemia, a bilirubin level above 15 mg/dl and any degree of 1653166 Death is most often in hepatic en~ephalopathy.~’, coma, with or without gastrointestinal hem~rrhage.~’ The terminal phase often includes renal failure.49 Correlations between clinical manifestations and morphologic changes are imperfect. The traditional view that the Mallory body is a histologic trademark of alcoholic hepatitis, both essential and pathognomonic for the diagnosis, has undergone r e v i s i ~ n . ~ ’ - ~Th ~ >e syndrome of alcoholic hepatitis (jaundice, fever, and leucocytosis) may be found in patients whose livers show fat and necrosis but few or no Mallory bodies.152Conversely, some patients may show the histologic features of alcoholic hepatitis despite the absence of the jaundice and fever considered characteristic of the syndrome.1s’,152 Indeed, the clinical picture in patients with the histologic features of alcoholic hepatitis can range in severity from a mild almost asymptomatic condition to a fatal syndrome of liver failure. Furthermore, the MB may be seen in other forms of liver disease (Table 4). Also recipients of 4,4’-diethylaminoethoxyhexestrol,’68 perhexiline maleate,98, amiodarone,94-96, 170 171 nifedipine, loo diethylstilbestrol,l o ’ and prolonged glucocorticoid treatrnentl7*may develop MBs. The MB has been reproduced in the mouse by administration of griseofulvin, 3,5-diethoxycarbonyl-1,4 dihydrocollidine,174and dieldrin,’75and in several strains of hamster by diethylstilbestrol.17‘ Administration of alcohol to experimental animals was at one time reportedI7’ to produce alcoholic hepatitis, but other efforts to produce Mallory Table 4. Some Agents Whose Hepatotoxtc Effects Are Enhanced in Animals by Alcohol Pretreatment or in Alcoholrc Humans

Acetaminophen Aflatoxin E Ally1 alcohol Eromobenzene Carbon tetrachloride Chloroform Cocaine Oiethylnitrosamtne Dimethylnitrosamine Enflurane

Galactosamtne Halothane lsoniazid Trrchlorobromoethane 1.1 ,2,-Trichloroethane Trichloroethylene Vinyl chloride Vinylidene chloride Vitamin A

54

bodies in experimental animals by administration of alcohol have not been successful. The significance of alcoholic hyaline remains somewhat uncertain. It is clearly not as specific for the entity of alcoholic liver disease as once thought. Nevertheless, it is a characteristic lesion of alcoholic hepatitis, to some degree a measure of severity of the injury in that condition, and, presumably, an index of a type of injury, if not of a specific etiology. That it reflects a toxic injury may be deduced from the production of the MB by administration of a number of chemicals and drugs and from the observation that an apparently identical cellular lesion is seen in the pneumocytes of asbestosis.I7*

Cirrhosis As in other stages of ALD, the clinical picture is the resultant of the three vectors: portal hypertension, loss of parenchymal function, and extrahepatic injury produced by alcoholism. In advanced cirrhosis, the chief manifestations result from portal hypertension, expressed as esophageal varices, splenomegaly,and hypersplenism. The distorted portal venous flow, coupled with inadequate parenchymal hepatic function, is also responsible for the easily precipitated hepatic encephalopathy and for the development of ascites. Jaundice is not prominent in cirrhosis unless there is some degree of alcoholic hepatitis. The chief manifestation of loss of parenchymal function is hypoalbuminemia. Death is most often the result of bleeding esophageal varices, hepatic coma, and the hepatorenal syndrome, usually in combination; infection, or other intercurrent illness may be responsible for death.21,132 In the patient with ascites, bacterial peritonitis is an easily overlooked and frequently fatal complicati011.I~~ Hemochromatosis An intimate relationship between alcoholism and hemochromatosis has interested investigators since the early years of this century;'79 and until the classic monograph of Sheldon'*' appeared, suggesting that the condition was genetic, there were many authors who considered hemochromatosis to be a form of alcoholic cirrhosis. Indeed, iron overload is prevalent among alcoholics with liver disease. Forty to 50% of those with fatty liver or cirrhosis have excess iron stores in the liver,181,185 and in a few percentages the degree of iron overload, evaluated histochemically, may seem as great as that seen in constitutional hemochromatosis. Is' Measurement of iron content, however, serves to distinguish the alcohol-associated iron overload from constitutional hemochromatosis. I** "Secondary hemochromatosis" had been considered to present a syndrome difficult to distinguish from that of the constitutional variety,'83and the "acquired hemochromatosis" of alcoholism had seemed much more prevalent than the genetic variety.Is4It is the current view, however,

ISHAK ET AL.

that alcoholics who develop the full syndrome and SUEcient iron overload have the genetic defect for hemochromatosis, the effect of which is potentiated by the alcoholism.IR5Indeed, the high prevalence of the abnormal gene with the frequency of heterozygotes ranging from 8% to 14% and of homozygotes from 0.25% to 0.5%182,185 demonstrates that the condition is not rare. Nevertheless, the great prevalence of striking iron overload in South African blacks who consume large amounts of iron and alcohol1s5 indicates that iron overload approaching that of genetic hemochromatosis may be secondary to alcoholism when there is also very heavy iron intake. Factors that might be responsible for iron overload in the alcoholic include: (1) increased intake and absorption of iron; (2) impaired utilization of iron due to the impaired erythropoiesis of folate deficiency and of the myelotoxicity of alcohol; (3) repeated bursts of hemolysis; and, perhaps, (4)enhancement of iron deposition as the result of liver damage. Portosystemic shunting apparently can provoke iron deposition in the liver,Is6 but the effect is variable and often trivial. Clinical features may not distinguish alcoholic cirrhosis without iron overload from that with, or indeed, from constitutional hemochromatosis. The syndrome of cirrhosis, in both, may be accompanied by diabetes, hypogonadism, and skin pigmentati~n.'*~,'*~ The prognosis of the hemochromatosis accompanied by alcoholism also appears to be that of the ALD. Only the far higher incidence of hepatocellular carcinoma distinguishes the outlook of hemochromatosis where it approaches 25%185 from that of alcoholic cirrhosis where it is 5% to 15%.

Porphyria Cutanea Tarda (PCT) This chronic syndrome is characterized mainly by bullous skin lesions on exposure to sunlight, liver disease, and iron o ~ e r l o a d . ' The ~ ~ ~entity ' ~ ~ has a number of features in common with hemochromatosis. Both are genetic conditions, and for both there is also an acquired form that can be provoked or enhanced by alcoholi~rn.'~~ Both may be accompanied by alcoholic liver di~ease."~~Both are characterized by excess stores of ironIs7and are clinically benefited by reducing the iron stores by p h l e b o t ~ m y . ' ~ ~ ~ In both the incidence of hepatocellular carcinoma is high.'85,190 Prominent differences are the extrahepatic organ involvement in hemochromatosis, of which there is little in PCT, and the greater iron stores in hemochromatosis than in PCT. Particularly pertinent to this discussion is the associatibn of PCT with alcoholism. Over 50% of individuals with PCT consume alcohol in excess.6' The frequency of PCT among alcoholics, however, is very low. There are also nonalcoholics with this condition.'" Some had been taking estrogens, oral contraceptives, or other drugs or had been exposed to other chemical^.'^' These may include instances of genetic PCT brought to light by the effects of the drugs or alcohol. There is, however, no doubt

ALCOHOLIC LIVER DISEASE

that exogenous toxic agents can lead to a PCT syndrome independent of a genetic predisposition, as was demonstrated clearly by the huge epidemic of cutaneous porphyria in association with severe liver disease among Kurds in Turkey who had ingested wheat treated with the fungicide hexachlorobenzene.192 Biochemical features of PCT include increased porphytins in the liver, a high level of uroporphyrin in urine and feces and, to a lesser degree, increased coproporphyrin in the feces.’93Biochemical abnormalities may reflect the fatty liver or cirrhosis that is often present. The relationship between alcoholism and iron overload that applies to both porphyria cutanea tarda and hemochromatosis is intriguing. It is also of interest that the secondary hemochromatosis of the South African black is often accompanied by PCT.’” The biochemical lesion that leads to the accumulation of porphyrins in genetic PCT appears to be deficient activity of uroporphyrin de~arboxy1ase.I~~ The manner in which alcohol or other exogenous agents act to provoke the syndrome is not clear. PATHOGENESIS OF ALD

Today, there is little doubt that ALD is mainly a consequence of the toxic effects of alcohol.’95The view that prevailed during much of this century attributing the liver disease to the malnutrition often accompanying alcoholism has been largely set aside; although the possibilities of a contributory role of malnutrition in the pathogenesis of ALD has not been d i ~ c a r d e d . l ~ ~ - ~ ” Evidence for the hepatotoxicity of ethanol derives from epidemiologic and experimental data. Alcoholism remains the most important apparent cause of cirrhosis, especially in the Western World.200Epidemiological studies have demonstrated a relationship between regular intake of ethanol, amount taken and duration of use and the incidence of cirrhosis.20’ In support of these observations, experimental studies have demonstrated hepatotoxic effects of ethanol.Ig5 Intake of 80 g of alcohol per day (roughly equivalent to a quarter-liter of whiskey or 1 liter of wine) has been defined as “hazardous drinking”;*02since, with consumption of amounts in excess of that, the risk of developing cirrhosis increases ~ignificantly.~’~ Progressive increases beyond this figure are associated with progressive increases in prevalence and greater severity of liver disease.201 Lelbach2” has demonstrated an even more significant correlation between consumption of ethanol and development of cirrhosis by inclusion of the duration of the consumption. He found the prevalence of cirrhosis to rise progressively from a low of 8% of patients who had been taking a mean of 177 g per day for 8 years to a high of 5 1% for those who had been taking 227 g per day for over 20 years. Indeed, he has attempted to refine the correlation further by relating the prevalence of cirrhosis to the esti-

55

mated total amount of alcohol consumed during the “drinking lifetime,” expressed as g/kg of body weight. S o r e n ~ o nhowever, , ~ ~ ~ has suggested that, beyond a threshold, there is no further influence of amount consumed on the development of cirrhosis. The role of alcohol in causing cirrhosis, according to him, may be permissive rather than dose-related. Nevertheless, the available epidemiological data and experimental evidence of hepatotoxic effects strongly support a toxic role for ethanol in the production of cirrhosis. Host factors also must be important in determining susceptibility to the hepatic disease of alcoholism. Many alcoholics do not develop cirrhosis. Constitutional differences in vulnerability or exposure to other liver injury presumably account for the range of 10% to 50% of alcoholics who develop cirrhosis. Constitutional markers of susceptibility to alcoholic liver injury have long been sought. Among possible endocrine or metabolic differences that might affect susceptibility, however, only gender is clearly relevant. Contrary to traditional views, females are distinctly more susceptible than males;’05 and the estimated hazardous level of intake for females is as low as 40 to 60 g/da~.~O’ A possible explanation for the greater susceptibility of females has been provided by studies on gastric alcohol dehydrogenase.206The “gastric barrier” to alcohol, provided by gastric metabolism of up to 20% of ingested ethanol, is less effective in women because of lower levels of activity of alcohol dehydrogenase in the stomach.z06The increased amount of alcohol available for absorption permits a larger proportion of each dose to reach the liver. Genetic differences in susceptibility to ALD among alcoholics has long been the subject of speculation, but there has been little evidence to support the likelihood. A genetic influence on the development of alcoholism has been clearly even an incriminable gene has been identified.208Efforts to identify genetic markers for susceptibility to ALD, however, have been inconclusive. Examination of distribution of HLA antigens in various studies of patients with ALD has revealed an increased prevalence of B5, B8, B13, B15, B40, Cw3, DR2, DR3, DRw9; and a decreased prevalence of B13 and B40 has been deClearly, among so large a number of incriminated HLA antigens, relevant markers, if there are any, cannot be identified. On the positive side, however, a gene controlling fibrogenesis (type 1 collagen gene) has been found to be more frequent in patients with alcoholic cirrhosis than in alcoholics without cirrhosis.210Extension of these studies may identify a genetic marker of susceptibility. Immunological responses affecting development of ALD have been the subject of many studies, and there is now considerable evidence of altered humoral and cellular immunity.209Evidence for abnormal humoral response in patients with ALD includes hypergammaglobulinemia involving IgA, IgC, and IgM, and circulating antibodies to

56

liver cell antigens or ne~antigens.~’~ Altered cell-mediated immunity has been inferred from the decrease in circulating T cells,211,212 attributed to their sequestration in the liver where they appear to be activated.*09There is also evidence of impaired T-cell regulation with inability to generate suppressor T-cells in alcoholic hepatitis;*I3and T cells or K cells cytotoxic for hepatocytes have been demonstrated in the Additionally, the demonstration that the surface of hepatocytes of alcoholic hepatitis bears HLA class I antigens that serve for attachment of sensitized T cells with the same antigen, serves to support a role for cellular immunity in production of ALD.209 Antigens that have been incriminated in triggering the immune humoral or cellular response include Mallory bodies or derived proteins, Mallory body material complexed to other proteins, hepatocyte protein dubbed “liver specific protein,” liver membrane proteins and a neoantigen apparently produced by action of ethanol or acetaldehyde on hepatocyte rnernbrane~.~’~ Of particular interest are recent reports of acetaldehyde adducts serving as neoantigens.*09 Putative mechanisms for injury have included deposition of immune complexes of antibody and one of the possible antigens on hepatocytes, and attack by activated T cells or K cells.*09The manner in which immunological factors would interface with toxicity of ethanol in provoking ALD is not clear. MacSween and Anthony209suggest that the initial injury leading to steatosis results from the toxicity of alcohol; and that the progression to alcoholic hepatitis and cirrhosis in some individuals may be the result of immunological events. This remains hypothetical, and it is possible that the immunological developments and markers in ALD are epiphenomenal. An immunologic factor in the clinical manifestations and perhaps pathogenesis of ALD is tumor necrosis factor (TNF).209aBird et al.209ahave found that plasma levels of TNF were elevated in patients with alcoholic hepatitis and that the degree of elevation appeared to correlate with the seventy of the hepatic disease. While the correlation is impressive, the data do not permit distinction between a pathogenic role for TNF elevated levels as a result of alcoholic hepatic injury. A critically important series of studies by Lieber and his colleagues during the past quarter-century have shown conclusively that ethanol per se can lead to a fatty liver in experimental animals and in man and to cirrhosis in the baboon. 19’ Indeed all of the histological features of alcoholic liver disease, except alcoholic hyaline, have been convincingly reproduced in animals by chronic administration of Lack of reproduction of alcoholic hyaline, a key feature of alcoholic liver disease, is a striking omission; since several drugs and other chemicals (see earlier pages) can reproduce alcoholic hyaline in experimental animals, yet alcohol does not. The striking success of Lieber and DeCarli in producing hepatic steatosis by the feeding of alcohol derives from the

ISHAK ET AL.

incorporation of the ethanol into a totally liquid diet.216 By this innovative approach, they were able to overcome the aversion of rats to alcohol and to induce them to accept 36% of the total calories consumed in the form of ethanol. They also have shown that ethanol can lead in a matter of days to hepatic steatosis in primates, and that alcoholic beverages taken in subintoxicating doses for a few days can lead to hepatic steatosis in humans, despite provision of an adequate or even enriched diet.’95Reproduction of necrosis and cirrhosis in the baboon has been reported in the baboon by Lieber and DeCarli.*16Nutritional factors in the pathogenesis of ALD continue to demand attention, and a contributory role of abnormal nutrition to the production of ALD cannot be dismissed. 1y7-200 Dietary manipulations can modify the effects of ethanol. Some studies have shown that a choline- or proteindeficient diet can enhance the adverse effects.’95s217,218 The fat content of the diet also can affect the degree of steatosis that is induced by ethanol, a low intake decreasing and a high intake increasing it.195,2’9 Furthermore, the greater degree of steatosis resulting from the high fat diet appears to enhance the development of necrosis and cirrhosis in the alcohol-fed rat.*I9The view that ethanol can produce severe steatosis in the rat and cirrhosis in the baboon, despite provision of a “normal diet,” has been challenged, however, by reports that more adequate nutrition can prevent the respective injury. 98-200, 18, Nevertheless, the accumulated studies in man and experimental animals leaves the view that the hepatotoxic effects of alcohol are importantly responsible for liver disease in the alcoholic; although nutritional abnormalities may well contribute to the hepatic disease. An effect of differences in nutritional status on the development of liver disease among alcoholics, however, remains to be demonstrated. The possible potentiating toxicity of other agents in the environment or the diet on development of ALD also warrants con~ideration;’~’**~~ but there are no data demonstrating that differences in exposure of alcoholics to other toxins explains the overall incidence of ALD or differences in susceptibility. A role of coincidental viral hepatitis in the pathogenesis of cirrhosis in the alcoholic has been considered although challenged.226Indeed, while the possible role of hepatitis B infection has seemed somewhat controversial,226recent evidence that the incidence of hepatitis C infection is far higher among alcoholics with liver disease than among other alcoholics2*’ suggests that this readily overlooked infection may play a role in pathogenesis of liver disease among alcoholics. Nevertheless, the apparent toxicity of ethanol, deduced from epidemiologic studies, has found equally compelling support in the experimental studies.

’

Pathogenesis of Individual Lesions of ALD Hepatic steatosis can be attributed to the introduction of physiologic lesions at multiple sites of lipid metabolism.

ALCOHOLIC LIVER DISEASE

The manner in which necrosis, alcoholic hyaline, and cirrhosis are produced is more difficult to explain.

Hepatic Steatosis Theoretically, fat can accumulate in the hepatocyte as the result of increase in the rate of delivery of lipids to the hepatocyte or in the synthesis of lipids by the hepatocyte, or decrease in oxidation of fatty acids by the liver or in egress of lipids from the liver.’33The fatty liver produced by most steatogenic agents results mainly from interference with egress of lipid.’33Ethanol, however, can provoke each of the mechanisms that could lead to increase in the lipid content of the liver. Most important in the steatogenic effect is impairment of disposition. 195 Increased synthesis of lipid, earlier considered important in alcoholic steatogenesis, now appears to be of little moment. Early studies suggested that the increased NADH/NAD ratio secondary to the oxidation of ethanol and of acetaldehyde enhances lipogenesis; but more recent studies suggest that increased synthesis of fatty acids is of little importance in the accumulation of fat.195The increased NADH/NAD ratio, however, may lead to enhanced synthesis of triglyceride by enhancing concentrations of alpha-glycerophosphate. The glycerophosphate, by trapping fatty acids reaching the liver or synthesized there, leads to increased triglyceride formation.’95 Increased delivery of fatty acids to the liver from the depots can contribute to hepatic steatosis, and variations in the amount of lipids absorbed from the gut can modify the development of hepatic s t e a t ~ s i s ;but ’~~ these ~ ~ fac~~ tors are apparently also of minor importance. Zncreased mobilization of lipid from peripheral depots is provoked by large, intoxicating doses of ethanol. ‘95,229 Presumably this results from the increase in circulating catecholamines provoked by the stressful intoxicating doses. The phenomenon does not occur with regular, smaller doses.229Furthermore, Lieber has discounted the relevance of the mobilization of peripheral lipids to the hepatic steatosis of chronic consumption of ethanol, since the lipids appear to be derived primarily from dietary rather than peripheral sources.229Nevertheless, increased mobilization of tissue lipids perhaps cannot be so readily dismissed in clinical states, since many alcoholics indulge not only in regular, chronic consumption of alcohol but also, intermittently, in larger, acutely intoxicating doses. Decrease in the rate of disposition contributes most importantly to the ~ t e a t o g e n e s i s . ’A~ ~decrease ’ ~ ~ ~ in rate of oxidation of fatty acids by mitochondria has been ascribed to the injury of that organelle produced by ethanol and acetaldehyde, as well as to alterations in mitochondria1 oxidative activity secondary to the NAD depletion engendered by the metabolism of ethanol. 195,229,230 Liebe?29 has emphasized the importance of the functional defect in oxidation of fatty acids that results from the effects of ethanol metabolism on the activity of the citric acid cycle, citing the metabolic block

51

produced by the increased NADH/NAD ratio at sites of the cycle requiring NAD (e.g., alpha-ketoglutarate oxidation). The resulting decrease in the generation of oxaloacetate, the pivotal molecule of the citric acid cycle, presumably leads to utilization by the mitochondria of the hydrogen equivalent from the ethanol oxidation rather than the acetyl fragments derived from oxidation of fatty The impaired mitochondria1 oxidation of lipid appears to be the physiologic defect most likely responsible for the hepatic steatosis induced by e t h a n 0 1 . ’ ~ ~ , ~ * ~ Decreased rate of egress of lipid from the liver has been to be of little importance in the lipid accumulation, at least early in alcohol-induced steatogenesis. This is in sharp contrast to the critical role of impaired exit of lipid from the liver in the genesis of the fatty liver caused by a number of other he pa tot ox in^.'^^ Indeed, ethanol administration has been reported to lead to increased rather than decreased release of lipoproteins to the Hyper-, rather than hypolipoproteinemia, characterizes clinical and experimental alcoholism, and secretion of the VLDL, the vehicle for egress of lipid from the liver has been described229to be high rather than low. Nevertheless, alcohol administration has been reported to lead to accumulation of an abnormal VLDL in the Golgi apparatus.z3oPerhaps the accumulation of fat induced by ethanol should be regarded in part as a form of “highoutput failure” of lipoprotein secretion. The rate of egress of lipid, though increased, is insufficient to cope with the increased rate with which triglyceride molecules are presented for transport. Furthermore, some interference with lipoprotein secretion or apoprotein synthesis has been reported to be induced by ethanol,231particularly in advanced alcoholic liver injury. Altered choline requirement was at one time considered to be a possible mechanism by which ethanol leads to hepatic ~ t e a t o s i s This . ~ ~ ~possibly has received relatively little attention in recent years and would seem to be belied by the differences between the effects of choline deficiency and of alcohol consumption on the liver. Choline deficiency leads to lowered phospholipid and carnitine levels in the liver,227while ethanol leads to increased levels of both.233Furthermore, even large doses of choline do not prevent or reverse the effects of ethanol; indeed, they enhance the toxicity.Ig5Presumably, the impaired uptake of choline by the liver, induced by is of little moment in producing the steatosis. Nevertheless, a contribution by nutritional factors to the hepatic steatosis of ALD cannot be dismissed. Protein and lipotrope deficiency can lead to fatty liver, especially in the growing animal and in children with kwashiorkor.133 The frequency of deficient protein intake among alcoholics and the reported therapeutic benefit of improved diet despite continued ethanol intake by alcoholics suggest that protein and lipotrope deficiency may play a role in the development of fatty liver and the severity of ALD.197,235

58

Hepatic Necrosis The genesis of necrosis is far less clear than that of steatosis. Action of ethanol as a "hepatocellular poison," or toxic effects on mitochondria of the acetaldehyde, produced by ethanol oxidation, has been incriminated230,236 as a cause of necrosis. Indeed, a necrogenic role for acetaldehyde, which is a very reactive molecule, seems a credible suggestion. It binds covalently to a number of macromolecules of the hepatocyte, a phenomenon that may be responsible for injury to mitochondria, plasma membranes, microtubules, and other organelles. Acetaldehyde also appears to provoke the lipid peroxidation of alcohol-induced hepatic i n j ~ r y . ' ~ ~ , ~ ~ ~ Peroxidation of membrane lipids has long been incriminated as a factor in ethanol-induced hepatic injury.195 Originally proposed as the explanation for hepatocyte injury permitting the accumulation of fat, peroxidative injury might also contribute to necrosis.237Evidence that ethanol leads to peroxidation in the liver has been provided by the evolution of malonyl dialdehyde and dienes and from the ability of antioxidants to inhibit the hepatic Proposed pathways for the peroxidative effects include the action of acetaldehyde, the microsomal generation of free radicals under the influence of e t h a n ~ l , ~and ~ ' , the ~ ~ enhanced ~ iron-stimulation of the peroxidation secondary to microsomal induction by ethan01.I.~~~ Another possible source of peroxidation is the bioactivation of other toxic agents whose metabolism is enhanced by the cytochrome P450 induction of ethan01.l'~ Also contributing to peroxidation is the reduction of hepatic glutathione (GSH) content, in part because of acetaldehyde binding to the molecule and in part because alcohol intake interferes with GSH synthesis.1-3,24' The depletion of GSH would enhance susceptibility to peroxidative injury.217The importance of these speculative peroxidative effects of ethanol in the production of liver injury, however, remains controversial and may apply only to acute effects of ethanol.240 A hypermetabolic state of the liver has been incriminated in ALD241-243 and considered responsible for the particular susceptibility of the alcohol-exposed liver to hypoxia. Treatment of rat liver in vitro has been shown to increase oxygen consumption, perhaps due to induction of cytochrome P450, to uncoupling of mitochondria1 respiration, or even to enhanced thyroid activity. Increased susceptibility to anoxia, accordingly, has been offered as an explanation for ethanol-induced necrosis. Ethanol leads to enhanced susceptibility to hepatic necrosis on exposure to hypoxia, similar to that of hyperthyroid animals. Furthermore, the recognition that the main brunt of alcohol-induced hepatic injury involves zone 3 supports this intriguing hypothesis. Steatosis begins in zone 3 , alcohol hyaline is most prominent in zone 3 , early fibrosis is mainly in zone 3 , and the severe sclerosing fibrosis that develops in some patients with alcoholic hepatitis is also centrilobular. However, the production of acetaldehyde

ISHAK ET AL.

by the cytochrome P450 system, which is also concentrated in zone 3, offers an alternative explanation for the zone 3 location of the brunt of alcohol-induced hepatic injury. 195 (The microsomal capacity to metabolize ethanol has been termed the microsomal ethanol oxidizing system or MEOS.'95 MEOS is now recognized to include or be the P450IIEI isozyme.) The studies of the Toronto group have led them to attempt treatment of alcoholic hepatitis The reports of benefit245,246 with propylthi~uracil.~~~-~~~ are however, and the issue remains to be resolved.

Cirrhosis The cirrhosis has been considered to be the consequence of necrosis (of alcoholic hepatitis) rather than of alcoholic steatosis. Nevertheless, while there is much to support the cirrhotogenic role of alcoholic hepatitis, the role of steatosis in leading to cirrhosis may have been dismissed too readily.l41,248,249 Reason to accept the likelihood that al; coholic steatosis can progress to cirrhosis without an intervening alcoholic hepatitis stage has been cited on an earlier page. Perivenular and perisinusoidal fibrosis may develop in the steatotic liver as the harbinger of cirrhos~s.'~~ The increase of connective that is the dominant feature of cirrhosis appears to be the result of increased synthesis, decreased destruction of collagen or both.249 The demonstration that the liver contains enzyme systems for collagen synthesis and breakdown permits the supposition. Evidence for increased fibrogenesis in alcoholic liver disease includes biochemical markers in the blood, liver, and urine. Patients with alcoholic hepatitis have increased levels of protocollagen and its key component hydroxyproline in the blood, increased activity of proline hydroxylase in the liver, and increased levels of hydroxyproline in the urine.249While necrosis is credited with the main role as the trigger of fibrogenesis, the metabolism of ethanol also may contribute to fibrogenesis. The decreased pyruvate and increased lactate levels, resulting from the altered redox potential, have each been considered candidates for the stimulus of fibrogenesis by generating increased amounts of proline. Myofibroblasts, the cells apparently responsible for fibrosis, are derived from Ito cells, the fat storing cells. Increase in number of myofibroblasts appears to precede the appearance of fibrosis in the alcoholic fatty 1 i ~ e r . l ~ ~ Defective degradation of collagen may contribute to cirrhosis.249While there is no biochemical evidence that collagenolysis is impaired in alcoholic cirrhosis, the hypothetical possibility that dense or relatively vascular fibrosis might handicap collagenolysis has been considered.250 At least four vectors, accordingly, contribute to the production of the cirrhosis: ( 1 ) active fibrogenesis in response to necrosis and the metabolic stimuli already cited,

59

ALCOHOLIC LIVER DISEASE

(2) defective collagenolysis, (3) areas of collapse of reticulum secondary to necrosis, and (4) regenerating nodules that contribute to distortion of the architecture. The multiplicity of cirrhotogenic factors and variation in the stage in the life history of cirrhosis at the time of examination presumably are responsible for the variations in morphologic type of cirrhosis observed. The classical view that alcoholic cirrhosis is micronodular has been modified by the recognition that late cirrhosis may be macronodular or "mixed." ETHANOL AND DRUG METABOLISM

The interface between the metabolism of ethanol and that of other xenobiotics is of theoretical and practical impofiance.'-3,195,251-256 Ethanol is apparently metabolized mainly by the cytosol enzyme, alcohol dehydrogenase (ADH). A significant fraction, however, is metabolized by the cytochrome P450 system, particularly at high blood ethanol levels or on chronic intake.'95 Both the ADH and the MEOS pathways yield acetaldehyde, a toxic intermediate, which is ultimately oxidized to C02 and water. Ethanol can enhance its own metabolism apparently by induction of the P450 and the activity of the MEOS.'95 The inducing effects of ethanol can also be seen in the hyperplasia of the SER and in the enhanced metabolism of a number of drugs by alcoholic humans and by ethanoltreated rats.'-3,251-257 By the same token, ethanol while still in the liver can competitively inhibit the metabolism of other The inducing effects of ethanol in humans disappear after 3 weeks of abstinence,257and in alcoholic patients who have developed severe liver disease.258Indeed, alcoholics with overt hepatic disease have shown subnormal ability to metabolize some xenobi~tics.~'~ These phenomena shed light on important clinical events. The effects of some drugs (e.g., CNS depressants) are enhanced by the simultaneous intake of ethanol because of "competition" for m e t a b ~ l i s m .They ' ~ ~ may be enhanced also in patients with severe hepatic disease.237 Conversely, alcoholics without severe liver disease may show increased tolerance for drugs as a result of the induction by alcohol of the cytochrome P450 Adaptive responses to ethanol also have a bearing on hepatotoxic phenomena. Induction of the P450 system could, of course, decrease the toxicity of agents that are inactivated by that system and enhance the toxicity of those that are converted to toxic metabolites. Indeed, the long-recognized, increased susceptibility of alcoholics to CC1, toxicity', ' 9 5 , 2 5 2 - 2 5 5 can be ascribed to the inducing effects of ethanol on the P450 and consequent enhanced conversion of CCl, to the active radical.'. ' 9 5 , 2 5 2 - 2 5 5 The enhanced hepatotoxic effects of CC14 in alcoholic patients has been known since the early years of this century.' First inferred from anecdotal observations, it has since been

amply confirmed in experimental a n i r n a l ~ . ~ ~A~variety -*'~ of explanations for the phenomenon were offered, but none sufficed.' It was only with the recognition that toxic agents were transformed to toxic metabolites by the cytochrome P450 system, and the recognition that ethanol enhances the activity of the P450 system that the basis for ethanol enhancement of a number of hepatotoxic events became clear. The induction of the cytochrome P450 system reflects a very important effect of alcohol on the liver. It explains the enhancement in alcoholics of the adverse effects on the liver of CC14, CHCI3, and of a number of other halo alkane^,^^^ probably explains the increased adverse effects of vinyl chloride on the liver,259and, indeed, the enhanced adverse effects of a number of hepatotoxic agents or drugs (Table 4). Of particular recent interest has been the enhancement by alcoholism of the hepatotoxic effects of acetaminophen (ACM).'-3,'95,255-259,260 While in the nonalcoholic individual the hepatotoxic dose of ACM usually exceeds 15 g and always is in excess of 6 g, in the alcoholic even doses in the recommended therapeutic range (