| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Department of Pharmacy Practice and Administration, Rutgers University College of Pharmacy, PO Box 789 William Levine Hall, Piscataway, NJ 08855-0789.
2
Division of Pediatric Pharmacology and Toxicology,
University of Medicine and Dentistry of New Jersey–Robert Wood Johnson
Medical School, 681 Frelinghuysen Rd., PO Box 1179, Piscataway, NJ
08855-1179.
a Author for correspondence. Fax (908) 281-0706; e-mail glambert{at}umdnj.edu
 |
Abstract |
|---|
 Top Abstract IntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
Key Words: indexing terms: xenobiotic metabolism • cytochrome P450 enzymes • cytotoxic/cholestatic liver injury • liver function tests
|
Introduction |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
The liver is prone to xenobiotic-induced injury because of its central role in xenobiotic metabolism, its portal location within the circulation, and its anatomic and physiologic structure (4). The liver is divided into multiple lobules, each centered around a terminal hepatic (central) venule and surrounded peripherally by six portal triads. Afferent blood is supplied by the portal venules and hepatic arterioles of the portal triads, flows through the hepatic venous sinusoids, and empties into the terminal hepatic venule. The regional pattern of hepatocellular necrosis observed with some xenobiotic-induced liver injuries can be understood by dividing the liver into functional subunits referred to as acini (4)(5). Each liver acinus is divided into three concentric zones of hepatocytes radiating from a portal triad and terminating at one or more adjacent terminal hepatic venules. Hepatocytes closest to the portal triad (zone one) receive blood most enriched with oxygen and other nutrients and are most resistant to injury. Hepatocytes more distal to the blood supply receive a lower concentration of essential nutrients, making them more susceptible to ischemic or nutritional damage. Most important for xenobiotic-induced hepatic damage, the centrilobular (zone three) hepatocytes are the primary sites of cytochrome P450 enzyme activity, which frequently makes them most susceptible to xenobiotic-induced liver injury (6), as discussed below.
|
Contribution of Metabolism to Xenobiotic-Induced Liver Injury |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
Metabolic reactions are of two types, phase I and phase II (7)(8). Phase I (oxidation, reduction, or hydrolysis) reactions typically occur first, and enhance water solubility by generating hydroxyl, carboxy, or epoxide functional groups on the parent compound. These functional groups in turn facilitate phase II reactions, conjugation with glucuronate, sulfate, acetate, or glutathione moieties. Conjugation reactions generally serve to further enhance water solubility and renal excretion (7). Phase II reactions also play a role in the prevention of xenobiotic-induced liver injury because most conjugates are biologically inactive (7)(8)(9). Disruption of normal phase II processes can lead to accumulation of hepatotoxic phase I metabolites.
Phase I oxidation and reduction reactions are primarily catalyzed by cytochrome P450 enzymes, a supergene family of heme-containing, mixed-function oxidase enzymes found in greatest concentration in the smooth endoplasmic reticulum of centrilobular hepatocytes (7)(10)(11). These enzyme reactions have the potential to induce cellular injury via several mechanisms of toxicity. The cytochrome P450 enzyme-catalyzed oxidation of xenobiotics such as bromobenzene or acetaminophen generates a highly electrophilic intermediate capable of forming covalent adducts with critical cellular macromolecules such as thiol-containing membrane proteins that regulate calcium homeostasis (7)(8)(12). The induction of increased intracellular calcium concentrations may be the common pathway leading to cell death. Cytochrome P450 enzyme-mediated reduction of halogenated hydrocarbons such as carbon tetrachloride or halothane can also generate free radical intermediates, which can directly damage cell membranes via lipid peroxidation, or can target nucleophilic DNA residues (8)(13)(14)(15). Similar cellular damage can result from the generation of reactive oxygen species such as hydrogen peroxide and hydroxyl free radical during a process known as redox cycling (8)(16). Redox cycling occurs when a reduced quinone substrate such as menadione or doxorubicin spontaneously reoxidizes in the presence of oxygen, thereby reducing the oxygen molecule (17).
|
Determinants of Host Susceptibility to Xenobiotic-Induced Liver Injury |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
(1)(18). Intrinsic injuries are
predictable, in that a threshold dose exists in all individuals,
typically leading to zonal liver cell necrosis accompanied by little or
no signs of inflammation. These injuries are generally the result of
phase I bioactivation reactions, with damage mediated by reactive drug
metabolites as previously discussed. In contrast, the nature of
idiosyncratic liver injuries suggests that most of these are mediated
by an immune mechanism (18)(19).
|
Idiosyncratic liver injuries are usually associated with classic signs of hypersensitivity, including fever or rash, and liver biopsy specimens reveal evidence of monocytic or eosinophilic infiltrates (9)(20). These reactions tend to occur only after repeated exposure, suggesting the need for initial sensitization, and drug rechallenge generally elicits prompt reappearance of symptoms. Both humoral and cellular immune mechanisms have been implicated in these types of liver injuries (1). One proposed explanation is the formation of a metabolite that is either haptogenic or alters an hepatocyte macromolecule to generate a neoantigen (19). The idiosyncratic hepatotoxicity of nonsteroidal antiinflammatory drugs such as diclofenac have been linked to such a mechanism (21). The phase II glucuronidation of these compounds can also produce reactive acyl glucuronides, which may bind irreversibly to nucleophilic amino acid side chains in hepatocyte membranes, potentially inducing a cell-mediated or humoral immune response (21). T lymphocytes or immunoglobulin molecules targeted against a variety of presumed neoantigens have been identified. Some of these immunoglobulin molecules recognize the cytochrome P450 isoenzyme responsible for metabolism of the offending drug compound. Examples of anti-microsomal antibodies include anti-liver/kidney microsome (LKM)1 antibodies directed against cytochrome P450IID6 isoenzymes (anti-LKM1 antibodies), and anti-LKM2 antibodies, which recognize P450IIC9 (22)(23). Anti-microsomal antibodies targeting P450IA2 and P450IIIA1 isoenzymes have also been reported in idiosyncratic liver injuries associated with dihydralazine and aromatic anticonvulsants, as well as anti-LKM3 antibodies directed against uridine diphosphate glucuronyltransferase (UDPGT) enzymes (24)(25)(26).
Idiosyncratic drug-induced liver injuries can also exhibit a regional pattern of cellular damage and necrosis or a mixed type of pattern. A centrilobular pattern is associated with the use of halothane and other volatile halogenated anesthetic agents (27). As would be expected with idiosyncratic injury, halothane hepatitis is non-dose-related and exceedingly rare (28); however, it exhibits characteristics of both idiosyncratic and intrinsic liver injury (29). Halothane is extensively metabolized by cytochrome P450 enzymes, with the corresponding oxygen tension present during these reactions determining whether oxidation or reduction pathways predominate (30)(31). Conditions of relative hypoxia favor reduction and the generation of various hepatotoxic phase I intermediates capable of forming protein adducts. This process takes place in all individuals exposed to halothane, but antibodies directed against these presumed neoantigens are detectable only in individuals with severe liver injuries, suggesting individual host susceptibility to injury (32). The neoantigens include adducts with proteins in the endoplasmic reticulum (reticuloplasmins) and the pyruvate dehydrogenase enzyme complex of the mitochondria (33)(34).
However, some agents classified as idiosyncratic hepatotoxins cause milder forms of dose-related liver damage that cannot be explained by an immune mechanism (1)(2). These injuries are generally not accompanied by the usual signs of hypersensitivity and are slow to reappear on drug rechallenge. In addition, the threshold dose required to induce liver injury with many intrinsic hepatotoxins displays a great deal of individual variability. These observations have led to use of the terms "immunologic" and "metabolic" idiosyncrasy (18). Metabolic idiosyncrasy appears to be the result of inherent individual variability in the activity of hepatic drug-metabolizing enzymes. A variety of host-related factors are thought to contribute to metabolic idiosyncrasy.
variability in phase i enzymatic activity
Three CYP gene families, designated CYP1,
CYP2, and CYP3, encode the cytochrome P450
enzymes that play the major role in human xenobiotic metabolism (Table 2
) (7)(10)(35).
Genetic, physiologic, pathophysiologic, and xenobiotic-induced factors
that affect cytochrome P450 enzyme activity may help to account for the
increased susceptibility of certain individuals to drug-induced liver
injury, as outlined in Table 3
(1)(8)(36)(37)(38)(39)(40)(41)(42)(43). As a
general rule, women are at increased risk of drug-induced liver
injuries, particularly chronic ones. Oral contraceptives are known
inducers of cytochrome P450 enzyme activity, whereas pregnancy has been
shown to induce certain isoenzymes, such as P450IIIA4, and inhibit
others (42)(43)(44)(45)(46). Cytochrome P450IA2 activity is
gender-related, with males consistently exhibiting higher enzyme
activity (47). However, parity may be an important
determinant of P450IA2 activity. Parous females who lactated appear to
exhibit activity similar to that of males (47). Genetic
polymorphisms, characterized by poor and extensive metabolizer
phenotypes, have been identified in the P450IIC18, P450IID6, P450IIE1,
and possibly the P450IIIA4 isoforms and can alter susceptibility to
xenobiotic-induced liver injury (Table 3
)
(11)(39)(40)(48). For
example, the risk of perhexiline maleate-induced liver injury is higher
in individuals with the P450IID6 poor-metabolizer phenotype
(49).
|
|
Many drug-induced liver injuries are clearly age-related (50)(51). The activity of some cytochrome P450 isoenzymes (such as P450IA2 and P450IID6) is reduced by ~70% in neonates, followed by a rapid increase in activity during the first few weeks to months after birth to an amount two- to threefold more (for P450IA2) than that of adults. The activity of other P450 isoforms, e.g., P450IIIA enzymes, can be higher in newborn infants than in adults, and certain P450IIIA isoforms are primarily expressed only in the developing fetus (52)(53)(54).
The classic example in which altered activity of a cytochrome P450 isoenzyme can increase the risk of liver injury is acetaminophen toxicity. Ordinarily, >90% of an acetaminophen dose undergoes phase II glucuronidation and sulfation, yielding inactive conjugates that are excreted in urine and bile (9)(55). About 5% of a dose is oxidized by cytochrome P450IIE1 isoenzymes, and to a lesser degree by other P450 isoenzymes, to the hepatotoxic intermediate N-acetyl-p-benzoquinone imine (NAPQI). Hepatocellular damage is ordinarily prevented by phase II glutathione conjugation, which converts NAPQI to the inactive metabolite mercapturic acid. Acute ingestion of >~10 g of acetaminophen saturates the normal glucuronidation and sulfation pathways, leading to increased production of NAPQI, which rapidly depletes available glutathione stores. The risk of damage is increased, and the threshold dose lowered, with concomitant use of compounds such as alcohol or phenobarbital that are capable of inducing P450IIE1 activity (56).
variability in phase ii activity
Perhaps equally important to host susceptibility is the functional
capacity of phase II detoxification pathways. The most common type of
phase II reaction is glucuronidation, where glucuronic acid is
transferred from uridine diphosphate glucuronic acid (UDPGA) to a drug
or phase I metabolite by the enzyme UDPGT (7). UDPGT
enzymes are produced by two gene families, UGT1 and
UGT2 (57). At least six isoenzymes are encoded by
UGT1 genes and four isoforms by UGT2
(58)(59). The potential for individual
variability is illustrated by the fact that inducing agents such as
phenobarbital do not affect the activity of these isoforms equally
(60). The capacity of the glucuronidation process can be
inhibited by the temporary depletion of available UDPGA stores by drugs
such as acetaminophen and chloramphenicol (61). Other
drugs, including naproxen, ethinyl estradiol, and certain
benzodiazepines have been shown to directly inhibit UDPGT enzyme
activity (62)(63). Age can also alter UDPGT
activity, which is low at birth but increases steadily to nearly adult
values by age 1–3 months (64). Nutritional deficiencies
are another potentially relevant cause of deficient UDPGA stores
(65).
Sulfation reactions catalyzed by three families of cytosolic sulfotransferase enzymes represent important detoxification pathways for alcohols and phase I intermediates containing phenol groups (7)(66). The efficiency of sulfation reactions can be compromised by temporary depletion of inorganic sulfate pools by ingestion of drugs such as salicylamide (67).
Glutathione conjugation is critical in preventing liver injury from several agents, including acetaminophen and bromobenzene epoxide, by acting as a free radical scavenger (7). Acetaminophen overdose causes liver injury secondary to the temporary depletion of glutathione stores in the liver. Administration of the antidote N-acetylcysteine prevents further injury by stimulating glutathione synthesis, thereby replenishing liver stores (68). Glutathione stores are also sensitive to fasting and alcohol ingestion and, as in most phase II pathways except sulfation, glutathione conjugating activity is depressed in neonates, even though glutathione transferase enzyme activities are apparently within the reference interval (69).
Amine or hydrazine-containing drugs or phase I metabolites are detoxified primarily by phase II acetylation reactions, catalyzed by cytosolic N-acetyltransferase (NAT) enzymes [7]. NAT-1 and NAT-2 represent the two gene families currently known to exist in the human liver (70). Polymorphism in NAT-2 results in the rapid or slow acetylator phenotype, which has been implicated in host susceptibility to liver damage by drugs such as isoniazid (71)(72). Isoniazid undergoes extensive NAT-2-catalyzed acetylation to acetylisoniazid, which is then hydroxylated by cytochrome P450 enzymes to the hepatotoxic intermediate acetylhydrazine, a metabolite capable of forming covalent cellular adducts (73). The risk of liver toxicity is higher in slow acetylators, in the elderly, and in association with concomitant use of cytochrome P450 inducers such as alcohol or rifampin (74).
|
Diagnosis of Xenobiotic-Induced Liver Injury |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
A key element in establishing the cause of injury is the temporal relationship between drug administration and appearance of symptoms, as well as the withdrawal of therapy and resolution (9). Because many drug-induced liver injuries are much more severe with reexposure, rechallenge with the suspected agent is generally not recommended. A complete medication history, including all prescribed and over-the-counter agents, as well as potential exposure to occupational or environmental chemicals, should be carefully documented and screened.
clinical, biochemical, and histologic patterns of
xenobiotic-induced liver injury
Xenobiotic-induced liver injuries can be broadly classified as
cytotoxic (necrotic or steatotic), cholestatic, or mixed (Table 4
) (1). The presence of an injury can be clearly
established on the basis of the clinical and biochemical evidence.
These indices can also be utilized in establishing a tentative
classification of the injury (76). However, histologic
examination of a liver biopsy specimen remains the only means of
definitively diagnosing the type of injury present (77).
|
Hepatocellular necrosis.
Clinically, hepatocellular
necrosis can range in severity from asymptomatic increases of
aminotransferase enzymes to jaundice to overt hepatic failure
(78). With intrinsic hepatotoxins, nonspecific
gastrointestinal symptoms such as nausea or vomiting typically begin
within a few hours of exposure. These symptoms often resolve within 48
to 72 h, followed by a 1- to 2-day period of relative well-being.
With severe injuries damage is ongoing, however, and biochemical
evidence (increased hepatic aminotransferase enzymes) of damage, often
accompanied by oliguria, begins to appear. Overt liver failure is
generally established within 3–5 days, characterized by jaundice,
coagulopathies, neurologic symptoms, and acute renal failure. The
degree of aminotransferase enzyme increase, hyperbilirubinemia, and
prolongation of the prothrombin time have prognostic significance, as
does the appearance of any manifestation of hepatic encephalopathy
(79)(80).
Toxic hepatitis.
Hepatocellular necrosis is a hallmark
of these injuries, but the associated symptoms and histologic pattern
of injury are nearly identical to those observed with acute viral
hepatitis, as outlined in Table 4
(1)(18)(81)(82)(83). Histologically,
these injuries typically reflect diffuse hepatocellular necrosis, which
may be associated with cholestasis. Lobular structure is generally
maintained, and even in severe cases, areas of necrosis are usually
surrounded by viable hepatocytes that reveal various degrees of
degenerative changes (18). Prominent monocytic or
eosinophilic inflammatory infiltrates are common (20).
These injuries are thought to result from bioactivation to toxic
metabolites (18).
Symptoms of toxic hepatitis range from asymptomatic increases of hepatic aminotransferase enzymes to signs of overt liver failure. With drugs such as phenytoin, these injuries often present with abrupt onset of fever and nausea, which may be accompanied by diffuse rash and arthralgias (1)(84). Evidence of severe liver injury (jaundice, coagulopathies, neurologic symptoms) appears 3–4 days after the onset of fever. The prodromal symptoms are usually mild or even absent with drugs such as isoniazid (85). As with hepatocellular necrosis induced by intrinsic hepatotoxins, clinical and biochemical markers have prognostic value (86). Extreme increases of hepatic aminotransferase enzymes, jaundice, and coagulopathies are associated with a mortality rate >10%.
Steatosis.
Steatosis results from the abnormal
accumulation of triglycerides within the hepatocyte, as summarized in
Table 4
(18)(87). Macrovesicular steatosis is
characterized by a single large cytoplasmic vacuole of triglyceride
within the hepatocyte that displaces the nucleus peripherally. The
etiology of macrovesicular steatosis is multifactorial, including
increased mobilization of fatty acids, increased hepatic synthesis of
fatty acids, increased synthesis of triglyceride from fatty acids, and
deficient removal of triglyceride from the hepatocyte via defective
VLDL synthesis (78)(88). Microvesicular
steatosis is a less common but more severe variant, resulting primarily
from deficient mitochondrial ß-oxidation of fatty acids and
characterized by the presence of multiple small droplets of
triglyceride within the hepatocyte, which do not displace the nucleus
(89)(90). The ß-oxidation of fatty acids is
a critical process, because the resulting acetyl-coenzyme A moieties
are the primary source of ATP in most cells. Disruption of this process
promotes the esterification of fatty acids in the cytoplasm to
triglyceride, robs the cell of energy, and leads to hyperammonemia via
inhibition of ureagenesis (91)(92).
Microvesicular steatosis can exhibit a diffuse or regional pattern and
in severe cases is accompanied by inflammation and hepatocellular
necrosis (1)(89).
Valproic acid is an established cause of microvesicular steatosis,
which resembles Reye syndrome and is in fact more likely to occur in
young children (93)(94). The lesion is
accompanied in severe cases by inflammation, necrosis, and bile duct
injury. Valproic acid-induced liver injury is thought to result from
phase I bioactivation (1)(95). Cytochrome P450
enzymes mediate the production of 
4-valproic acid, an
oxidative metabolite capable of generating coenzyme derivatives.
Production and accumulation of these derivatives may inhibit
mitochondrial ß-oxidation via depletion of free coenzyme A and
carnitine concentrations
(18)(89)(96). Early symptoms of
injury are insidious in nature, and include gastrointestinal complaints
and mental status changes; they usually appear during the first 4
months of drug exposure (94)(97).
A related condition termed phospholipidosis is produced by perhexiline maleate and the antiarrhythmic agent amiodarone (18)(98)(99). This dose-related liver injury is characterized by the accumulation of phospholipid–drug complexes within the lysosomes of hepatocytes, bile ductal cells, sinusoidal Kupffer cells, and various other tissues, including peripheral nerves, skeletal muscle, cornea, heart, lungs, and skin (100). The amphophilic nature of these agents is thought to promote the formation of drug–phospholipid complexes, called lysosomal inclusions. Both amiodarone and its N-desethyl metabolite readily form these complexes, subsequently promoting the accumulation of phospholipids via inhibition of normal phospholipase activity (101). These inclusions are present in virtually all patients treated with amiodarone, although overt liver injury is rare (102).
Cholestasis.
Xenobiotic-induced cholestasis
results from the disruption of bile production or flow and exhibits one
of two patterns, as outlined in Table 4
(1)(18). Hepatocanalicular (hypersensitivity)
cholestasis is characterized by prominent monocytic portal inflammation
and secondary damage to bile canaliculi, as seen with chlorpromazine.
Chlorpromazine and its 7,8-dihydroxy and 7-hydroxy metabolites
interfere with bile acid secretion via disruption of canalicular
membrane fluidity and Na+/K+-ATPase activity
(103)(104). Inhibition of phase II sulfation
pathways increases the risk of liver injury (105).
Evidence of immunologic idiosyncrasy includes the relatively rare
occurrence of this reaction (in <2% of the population), the rapidity
of onset (typically within the initial month of therapy), the prompt
reappearance of symptoms upon drug rechallenge, and the usual presence
of peripheral eosinophilia (18). Metabolic idiosyncrasy as
a cause of toxicity is supported by the observation that the sulfate
conjugate is not hepatotoxic, coupled with the fact that formation of
oxidative phase I metabolites would be expected to exhibit individual
variability. Nonspecific flu-like and gastrointestinal symptoms often
precede the development of jaundice, which is associated with intense
pruritis (18). Overt jaundice is generally accompanied by
extreme increases of alkaline phosphatase and conjugated serum
bilirubin. Hepatic aminotransferase enzymes are generally only mildly
increased in the absence of significant necrosis. Most patients recover
completely within 3 months of drug withdrawal.
In contrast, the use of estrogen or 17
-substituted steroids is
associated with canalicular or bland cholestasis, a less severe variant
(18)(106). Generally appearing within the
first 2 months of drug therapy, bland cholestasis is associated with an
insidious onset of jaundice and pruritis, and less commonly with
gastrointestinal complaints. Systemic signs of hypersensitivity are
absent. Bile stasis is most prominent in the canaliculi, without
accompanying inflammation or necrosis (107). The injury
typically resolves completely within 2 months of drug withdrawal.
Although asymptomatic mild increases of alkaline phosphatase or
aminotransferase enzymes are relatively common, jaundice is rare
(108). The influence of individual susceptibility is
illustrated by the fact that the risk of steroid jaundice is greater in
patients with a history of jaundice of pregnancy and in populations of
Scandinavian or Chilean descent (108)(109)(110).
Bland cholestasis appears to result from drug-induced alterations in
sinusoidal membrane fluidity and Na+/K+-ATPase
activity with no apparent effect on canalicular membranes, resulting in
a decreased hepatocellular uptake of bile acids from sinusoidal blood
(18). Host susceptibility may relate to differences in the
production and excretion of presumably nontoxic methylated metabolites
and glutathione activity (111).
Hepatic vascular injury.
Veno-occlusive disease is a
severe form of drug-induced liver injury characterized by thrombosis of
efferent hepatic venules, leading to centrilobular necrosis and liver
outflow obstruction, which can progress to congestive cirrhosis, as
outlined in Table 4
(18)(112)(113). The condition
presents clinically with abrupt onset of severe abdominal pain,
hepatomegaly, and jaundice, accompanied by extreme increases of hepatic
aminotransferase and alkaline phosphatase enzymes (114).
Veno-occlusive disease can progress rapidly to overt hepatic
failure, with manifestations of congestive cirrhosis such as ascites,
coagulopathies, and hepatic coma.
Oral contraceptive use is associated with hepatic venous thrombosis, a condition that can also lead to congestive cirrhosis resembling Budd–Chiari syndrome (16)(115). This injury is reportedly twice as common in oral contraceptive users as in nonusers and is thought to result from the thrombogenic properties of these agents (116). Oral contraceptives can also produce another type of vascular lesion called peliosis hepatitis, in which weakening of sinusoidal membranes leads to the development of blood-filled sacs within the hepatic parenchyma (117)(118).
Hepatic tumors.
Chronic use of oral
contraceptives is associated with the development of hepatic adenomas,
benign tumors typically observed only in women of childbearing age and
which were exceedingly rare before the widespread use of these agents,
as outlined in Table 4
(18)(119). These tumors
usually resolve completely with drug withdrawal, and risk of
development is highly correlated with duration of drug exposure
(120). Hepatocellular carcinomas have been associated with
the chronic use of anabolic, androgenic, and contraceptive steroids
(121)(122). The underlying mechanism by which
these agents produce tumors is not well understood, but malignant
tumors may be an extension of the same process, presenting initially as
benign adenomas (123)(124).
|
Classification of Injury and Evaluation of Residual Liver Function |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
)
(77).Each marker has inherent deficiencies in sensitivity or specificity,
and no single method appears capable of completely discerning the
etiology, severity, and prognosis associated with a given injury
(77)(126).
|
biochemical evaluation
Serum aminotransferase enzymes.
Serum activity
concentrations of aspartate aminotransferase (AST) and alanine
aminotransferase (ALT) are the most commonly used biochemical markers
of hepatocellular necrosis (77)(121). These
enzymes are localized in periportal hepatocytes, reflecting their role
in oxidative phosphorylation and gluconeogenesis. ALT is highly
specific for the liver, whereas AST is also located in the heart,
brain, kidney, and skeletal muscle, making this enzyme less specific
for liver injury (127). Serum activities in generally
healthy individuals are <0.58 µkat/L. These serum activities
presumably increase as a result of cellular membrane damage and leakage
(126).
Serum aminotransferase activities are increased in all types of hepatic injury, but they provide only a static estimate of the amount of recent damage and no indication of residual functional capacity (128). The highest increases (often >20-fold) are observed with acute hepatocellular injuries, such as xenobiotic-induced necrosis or acute viral hepatitis (129). Cholestasis or chronic liver disease rarely cause increases >8.3 µkat/L, and serum activities are generally within the reference interval or only slightly increased in alcoholic liver disease (130). The degree of increase does not correlate well with the extent of liver injury or prognosis (126). A decline in serum activity concentrations usually indicates recovery but in fulminant injury may be a poor prognostic sign, reflecting a major loss of functional hepatocytes (126)(127).
Serum alkaline phosphatase.
Four separate genes
encode this family of isoenzymes that catalyze the hydrolysis of
phosphate esters, generating inorganic phosphate
(127)(131). Sources of alkaline phosphatase
include the liver, bone, leukocytes, kidneys, and first-trimester
placenta. Serum values ordinarily range from 0.5 to 2 µkat/L but are
affected by several physiologic variables (132). Alkaline
phosphatase activities are markedly increased in children and
adolescents, as well as the third trimester of pregnancy
(133). In young to middle-aged adults serum activities are
usually higher in men; in elderly individuals the activities are often
higher in women (133).
Serum alkaline phosphatase increases to some extent in most types of liver injury. Bile acids account for this increase: They induce alkaline phosphatase synthesis and exert a detergent effect on the canalicular membrane, allowing leakage into serum (126)(134). Mild to moderate increases (less than threefold) are not specific for the type of liver injury (135). The highest concentrations are observed with cholestatic injuries (77). Alkaline phosphatase activity concentrations cannot be used to differentiate between intrahepatic or extrahepatic etiologies, because similar increases are observed with each type of bile stasis. The specificity of alkaline phosphatase for the liver is poor, for several other conditions (particularly bone diseases, growth spurts, or pregnancy) also increase serum values (77)(126).
Serum 
-glutamyltransferase (GGT).
GGT
enzymes are located in a variety of tissues, including the heart,
brain, kidney, pancreas, spleen, and the biliary ductule cells of the
liver (77). These enzymes catalyze both the transfer of
-glutamyl groups from peptides to amino acids and the metabolism of
glutathione conjugates. Serum activity concentrations are ordinarily
<0.5 µkat/L but may be higher in neonates and the elderly
(126). Serum concentrations correlate well with alkaline
phosphatase. The primary utility of monitoring serum GGT is in the
exclusion of bone disease as a cause of increased serum alkaline
phosphatase, a condition that does not affect GGT concentrations
(136). However, the enzyme is inducible by chronic alcohol
use and by drugs such as phenytoin (137).
Serum 5'-nucleotidase.
These enzymes are located in a
variety of tissues, including the brain, heart, blood vessels,
intestine, pancreas, and sinusoidal/canalicular membranes of the liver,
where they catalyze the hydrolysis of nucleotides, generating inorganic
phosphate (77). Serum activity concentrations ordinarily
range from 17 to 183 nkat/L and correlate well with serum alkaline
phosphatase concentrations (126). However, despite their
wide distribution, an increased serum value is highly specific for
cholestatic liver injury, given that the detergent action of bile acids
on the canalicular membrane is thought to be the only mechanism by
which the enzyme can gain access to the circulation (138).
The primary utility of 5'nucleotidase activities is in the diagnosis of
cholestatic liver injury in childhood or pregnancy, because neither
condition affects the serum values of this enzyme
(139)(140).
Serum bilirubin.
Free bilirubin is not water soluble and
must be bound to albumin to facilitate transport to the liver
(77). This indirect or unconjugated bilirubin fraction
therefore does not enter urine. Once taken up by hepatocytes, bilirubin
is conjugated by UDPGT enzymes, with the glucuronide conjugates (the
direct bilirubin fraction) being excreted in bile. Intestinal bacteria
metabolize direct bilirubin to urobilinogen, which is mainly excreted
in feces. A minor portion undergoes enterohepatic circulation, with
small quantities excreted in urine. Total serum bilirubin
concentrations (reference interval, 3–15 µmol/L) indicate the
functional transport capacity of the liver. The direct fraction
typically accounts for ~5% of the total serum value, and total
bilirubin concentrations are consistently higher in males.
Mild to moderate isolated, indirect hyperbilirubinemia is generally associated with hemolysis, although neonatal jaundice or inherited defects in hepatic uptake or conjugation can cause a similar pattern (77). Total bilirubin concentrations rarely exceed 70 µmol/L in such cases. Higher isolated increases or associated abnormalities of other liver enzymes indicate a hepatic etiology. When hepatic injury is present, the direct bilirubin fraction typically is at least 50% of the total serum value, but the total concentration rarely exceeds 500 µmol/L regardless of severity, because of renal excretion of the direct fraction. Urinary bilirubin is a more sensitive indicator of liver injury than is serum bilirubin. An increase in urinary bilirubin is nearly always indicative of a corresponding increase in the serum direct fraction attributable to intrahepatic or extrahepatic cholestasis. The degree of increase in serum bilirubin values has prognostic significance in chronic liver injuries, but not in acute injuries (141).
Serum bile acids.
Cholic acid and chenodeoxycholic acid
are the primary bile acids in humans (142). These organic
anions are synthesized in hepatocytes from cholesterol, conjugated to
glycine or taurine, and excreted into the canaliculus
(77). Measurement of serum bile acid concentrations is a
more specific indicator of functional hepatic excretory capacity than
serum bilirubin is (143). Serum bile acid concentrations
are also more sensitive to subtle excretory abnormalities, in that the
quantity of bile flow is so much greater than that of bilirubin.
An increase in serum bile acid concentrations in fasting is highly specific for liver injury and serves to exclude congenital or hemolytic causes of hyperbilirubinemia (144). The greatest increases are observed in acute viral hepatitis or extrahepatic cholestasis (77). The ratio of cholic to chenodeoxycholic acid (reference interval, 0.5–1.0) decreases with chronic injuries such as cirrhosis and increases with extrahepatic bile obstruction (145).
Serum albumin.
Serum albumin, the major plasma protein
synthesized in the human liver, is a clinically useful marker of
hepatic synthetic function (77). The relatively long
elimination half-life (~20 days) and ample storage pool, however,
limit the utility of this index in evaluation of chronic liver
injuries. In addition, several factors other than liver injury can
disrupt albumin synthesis, including nutritional deficiencies and
alterations in plasma oncotic pressure (hypergammaglobulinemia)
(146)(147). Alcoholic cirrhosis with or
without accompanying ascites generally lowers serum albumin
concentrations, although hypoalbuminemia in the setting of ascites may
reflect dilution rather than decreased synthesis
(126)(148).
Prothrombin time (PT).
PT provides an index of hepatic
synthetic capacity that applies to both acute and chronic liver
injuries (77). An indicator of the extrinsic clotting
cascade, the PT provides an indirect measure of the hepatic synthesis
of clotting factors I, II, V, VII, IX, and X (149). Other
causes of a prolongation of PT include vitamin K deficiency, warfarin
therapy, and acquired or congenital clotting factor deficiencies
(77)(150). The administration of vitamin K can
be helpful in distinguishing liver injury from vitamin K deficiency,
because a relative normalization of PT would be expected with the
latter etiology (which can result from obstructive jaundice), in
contrast to a lack of normalization when hepatocellular necrosis is
present (151).
PT has prognostic value in both acute and chronic liver injury. An extreme or worsening prolongation of the PT in the setting of acute hepatocellular necrosis is associated with an increased risk of fulminant injury (152). Similarly, a poor prognosis is associated with PT prolongation in the setting of chronic liver injury and after portal-systemic shunt surgery (153).
quantitative measures of liver function
Biochemical measurements provide a static assessment of the degree
of liver injury, but give little information about residual liver
function. In contrast, the administration of a compound metabolized by
the liver affords a dynamic evaluation of residual metabolic capacity
(functional liver cell mass) (77). To rule out confounding
factors such as alterations in hepatic blood flow or plasma protein
binding, suitable agents should have a low hepatic extraction ratio and
not be highly protein bound. Marker compounds should also be relatively
nontoxic and, if administered orally, should be both rapidly and
completely absorbed.
Caffeine clearance.
Caffeine elimination depends highly
on cytochrome P450IA2 isoenzyme-mediated N-demethylation,
which leads to a variety of urinary methylxanthine metabolites
(154). Several methods are used to assess liver function
with caffeine (155)(156). The
[13C]caffeine breath test is highly specific for P450IA2
isoenzyme activity, which catalyzes caffeine
3-N-demethylation (157). Breath samples are
collected over 1 h after oral administration, and the enzyme
activity is quantified via analysis of the ratio of
13CO2 to 12CO2 by
differential mass spectroscopy. Alternatively, caffeine or caffeine
metabolites (or both) can be analyzed in fasting plasma, urine, or
saliva by HPLC, but this also monitors metabolites that are produced by
several P450 isoforms, including P450IA2, P450IIIA4, and P450IIA6
(158)(159)(160).
Caffeine clearance is a reliable indicator of global hepatic function but is less sensitive for assessing mild injury (158)(160). It is most reliable in the assessment of cirrhosis but appears to be no better than the Child–Pugh classification scheme with respect to prognosis (161). Confounding factors may include smoking (P450IA2 isoenzyme induction), gender, age, or concomitant exposure to drugs that alter enzyme activity (162)(163)(164).
Galactose elimination.
Galactose is metabolized by
hepatic galactokinase enzymes rather than by the cytochrome P450 system
(77). Accordingly, less individual variability confounds
interpretation of test results (165). Multiple methods can
be used in measuring galactose elimination, including a
[14C]galactose breath test and intravenous administration
of nonradiolabeled drug, followed by serial blood sampling at 5-min
intervals between 20 and 45 min postdose
(166)(167). The accuracy of this approach may
be improved by eliminating the initial 20-min blood sample, which has
been shown to display the greatest amount of individual variability
(168).
Galactose elimination is defined by hepatic metabolism (thereby reflecting functional hepatocyte mass) so long as plasma galactose concentrations exceed 500 mg/L (169). At lower concentrations, hepatic blood flow becomes the primary determinant of clearance. Clearance is decreased by acute or chronic hepatocellular necrosis but is typically unaffected by cholestatic liver injury (170). However, most studies suggest that measurement of galactose elimination provides no prognostic advantage over the Child–Pugh classification (171). In addition, serum albumin concentrations appear to be just as effective as galactose clearance in distinguishing cirrhotic patients from healthy controls (166). In contrast, the redox tolerance test has been shown to accurately depict functional hepatic mass and provide prognostic information in patients with obstructive jaundice (172). Preoperative results correlate significantly (P <0.01) with serum bilirubin half-life. This test is based on an oral glucose load followed by 2-h serial arterial blood sampling to determine the cumulative enhancement of the ratio of acetoacetate to ß-hydroxybutyrate.
Formation of monoethylglycinexybutyrate (MEGX).
Lidocaine undergoes N-demethylation to MEGX via a phase I
reaction mediated by cytochrome P450 enzymes, including the P450IIIA4
isoform (77)(173). Serum concentrations of
MEGX are analyzed via fluorescence polarization immunoassay
(174). After baseline blood sampling and administration of
a single intravenous lidocaine dose, formation clearance can be
measured from serial blood samples obtained for 1–2 h postdose or a
single 15-min postdose MEGX concentration can be determined
(175). Factors that may confound test results include
gender (formation clearance is consistently higher in males), age (the
rate of MEGX formation declines with age), and concomitant exposure to
drugs that alter cytochrome P450IIIA4 activity (176).
Some studies suggest that MEGX formation has better prognostic value than the Child–Pugh classification in patients with end-stage liver disease, including selection of appropriate transplant candidates (177)(178). The serial use of MEGX formation clearance has also been shown to accurately predict hepatic failure in patients with multiple trauma, with the assessment on day 3 providing the greatest predictive value (179). However, other studies imply that MEGX formation clearance is less effective than galactose elimination capacity in assessing the severity of cytotoxic liver injury (180).
Antipyrine clearance.
The elimination of antipyrine
almost totally depends on metabolism by hepatic cytochrome P450 enzymes
(77)(181). In addition, antipyrine is nearly
completely absorbed after oral administration, is not bound to plasma
proteins, and is distributed into total body water, so that alterations
in protein binding do not interfere with clearance assessment.
Antipyrine clearance is measured after oral administration, followed by
collection of multiple blood samples.
Antipyrine clearance is most accurate in the assessment of chronic liver injury (182). In such instances the increase in elimination half-life correlates strongly with the degree of PT prolongation, hypoalbuminemia, and severity of hepatocellular damage (183). Antipyrine clearance is less sensitive in the assessment of acute liver injury. Limited evidence also suggests that antipyrine clearance may be a reliable predictor of clinical outcome in patients with obstructive jaundice (184).
The major problem with the use of antipyrine clearance in the evaluation of liver injury is the individual variability inherent with cytochrome P450 enzyme activity (185). As discussed previously, multiple patient-specific factors other than liver injury determine the metabolic capacity of the various cytochrome P450 isoforms. A "cocktail" approach with multiple markers, each primarily metabolized by a separate P450 isoform, has been proposed to overcome this problem (186)(187).
[14C
]Aminopyrine breath test.
This noninvasive test utilizes aminopyrine, an agent eliminated via
cytochrome P450 enzyme-catalyzed N-demethylation
(77). Serial breath sampling after oral administration of
14C-labeled aminopyrine permits assessment of metabolic
efficiency via analysis of the percentage of
14CO2 in expired air (188). A
single 2-h postdose breath sample appears to be equally accurate, with
presumably healthy subjects excreting 6.6% ± 1.3% of the dose by
this time.
The rate-limiting step in the formation of CO2 involves a folic acid-dependent enzyme. Several factors therefore affect aminopyrine metabolism, including folate or vitamin B12 deficiency, glutathione deficiency, protein deficiency, infection, and thyroid disease (189). Both acute and chronic hepatocellular injury are associated with a decrease in aminopyrine elimination, but the sensitivity of the test is poor, there being a great deal of overlap between healthy and diseased subjects (190)(191). Aminopyrine excretion generally remains within the reference interval in the presence of cholestatic injuries, regardless of etiology, but most studies imply little advantage of this approach over routine biochemical measurements in the differentiation of cholestatic from cytotoxic injury (192)(193). Limited evidence suggests that the aminopyrine breath test is a better predictor of survival in patients with alcoholic cirrhosis than are the Child–Turcotte and Child–Pugh classification schemes, although this conclusion has been disputed by other studies (194)(195)(196)(197).
|
Future Directions |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
Currently, the erythromycin breath test can monitor cytochrome P450IIIA activity in humans, and other substrates are under development to monitor specific liver enzymes (201). As more enzyme-specific substrates allow the in vivo examination of critical hepatic enzymes, the more we will learn about the roles of these enzymes in human development, birth defects, cancer, drug metabolism, and a host of human diseases.
|
|
Footnotes |
|---|
-glutamyltransferase; PT, prothrombin time; and MEGX, monoethylglycinexybutyrate. 
|
References |
|---|
|
TopAbstractIntroductionContribution of Metabolism to...Determinants of Host...Diagnosis of Xenobiotic-Induced...Classification of Injury and...Future DirectionsReferences |
|---|
-glutamyl cycle enzymes in human fetal liver. J Pharmacol Exp Ther 1981;217:697-700.
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  K. P DeSanty and C. M Amabile Antidepressant-Induced Liver Injury Ann. Pharmacother., July 1, 2007; 41(7): 1201 - 1211. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Shakya, B S. Rao, and B. Shrestha Incidence of Hepatotoxicity Due to Antitubercular Medicines and Assessment of Risk Factors Ann. Pharmacother., June 1, 2004; 38(6): 1074 - 1079. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. B. Duguay, I. M. Yousef, and G. L. Plaa Manganese-Bilirubin Effect on Cholesterol Accumulation in Rat Bile Canalicular Membranes Toxicol. Sci., January 1, 2000; 53(1): 150 - 155. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
