1092 www.thelancet.com/oncology Vol 9 November 2008

Review

Part I: Liver function in oncology: biochemistry and beyondKathryn M Field, Chris Dow, Michael Michael

The liver has a key role in the metabolism (ie, inactivation or activation) of many commonly used anticancer agents—cytotoxics or new biological agents. Therefore, assessment of liver function is a fundamental part of initial work-up and management of patients with cancer. An understanding of the meaning of conventional serum biochemical testing of liver function and status, what variables they are measuring, and usefulness for chemotherapy dosing is essential. Emerging awareness of the drawbacks of conventional serum biochemical testing and further understanding of the intricacies of liver function is leading to the development of alternative strategies for appropriate chemotherapy regimens and dosing. We present an overview of assessment of liver function and chemotherapy dosing. We consider the use of serum liver biochemical testing to predict liver function, potential causes of biochemical abnormalities in patients with cancer, and chemotherapy drugs that are associated with hepatotoxicity. Part II will overview the current knowledge surrounding chemotherapy dosing in the setting of liver dysfunction; as well as alternative tests of hepatic metabolic function that are beginning to be used as strategies for appropriate individualised chemotherapy administration.

IntroductionThe most appropriate strategy for eff ective and safe chemotherapy dosing in cancer is poorly defi ned. Patients with cancer have pharmacokinetic and pharmacodynamic variability, and many agents have a narrow therapeutic index. The astute clinician needs to be fully aware of these variables when selecting the most appropriate chemotherapy drug and dose for a patient.

Many cytotoxics are metabolised by the liver, causing drug inactivation (or activation if they are prodrugs). Chemotherapy in the setting of liver dysfunction can result in drug toxic eff ects (occasionally fatal), or can be associated with reduced eff ectiveness. Furthermore, several chemo-therapy agents induce liver injury or dysfunction, which can manifest as abnormal serum liver biochemistry or can be visualised histologically (eg, after neoadjuvant chemo-therapy before liver resection; fi gures 1 and 2).

Conventional serum liver biochemical testing does not always predict these potential complications. In many circumstances such testing is an indicator of liver infl ammation or damage, rather than of liver function. Although liver metastases seen on imaging might raise clinicians’ concerns about chemotherapy dosing, many other factors can impede eff ective drug metabolism by the liver that might be more diffi cult to recognise. We discuss current standard serum liver biochemical testing in clinical oncology, the measurements of liver physiology and pathophysiology they provide, and their limitations for chemotherapy dosing. We highlight the causes of abnormal serum liver biochemical tests in patients with cancer—treatment-related and non-treatment related.

Conventional liver-biochemical tests: adequacy for chemotherapy dosingSerum liver biochemical testing includes measures of synthetic function (ie, serum albumin and prothrombin time), and estimates of cellular injury (ie, aspartate aminotransferase, alanine aminotransferase), chole-stasis, or duct injury (ie, alkaline phosphatase, gamma-glutamyltransferase, and direct-reacting bilirubin).

The liver plays a major part not only in synthetic function, but also in metabolic and excretory functions, and conventional liver tests measure only a small proportion of the organ’s many functions. At best, liver tests are a small snapshot at the time blood is drawn of the current status of liver synthetic function, necroinfl ammatory activity, and biliary-tree integrity. Here, we use the term “liver tests” to describe conventional serum biochemical tests. Patterns of abnormalities are more meaningful than are elevations or reductions in individual tests: liver tests should be analysed together as a whole.

Serum concentrations of liver enzymes Activities of serum enzymes are commonly used to classify patients into cohorts of varying liver dysfunction for adjustment of chemotherapy dosing. However, they are not measures of liver function, synthetic capacity, or drug

Lancet Oncol 2008; 9: 1092–101

Division of Haematology and Medical Oncology, Peter

MacCallum Cancer Centre, Locked Bag 1, VIC, Australia

(K M Field MBBS Hons, M Michael MBBS Hons); and Department of Pathology, Royal Melbourne Hospital,

Victoria, Australia (C Dow FRCPA)

Correspondence to:Dr Michael Michael, Division of

Haematology and Medical Oncology, Peter MacCallum

Cancer Centre, Locked Bag 1, VIC 8006, Australia

Michael.Michael@petermac.org

Figure 1: Histology of healthy liverPortal tract contains bile duct, hepatic artery, and portal vein (upper right); central vein shown (lower left); and hepatocytes line sinusoids in centre of image. Cords and plates of hepatocytes are one-cell thick between sinusoids. Haematoxylin and eosin stain, ×200 magnifi cation.

www.thelancet.com/oncology Vol 9 November 2008 1093

Review

metabolism. Therefore, their inclusion in what are termed “liver-function tests” must be appreciated as a misnomer.

The liver contains many enzymes, four of which are commonly measured. Aspartate aminotransferase and alanine aminotransferase are usually intracellular enzymes, but are released when hepatocytes are damaged. Aspartate aminotransferase, also called serum glutamate oxaloacetate transaminase, catalyses the reversible transfer of an aminoacid group to the α or two-carbon atom of a ketoacid. For aspartate aminotransferase, the transfer of the amino group from aspartic acid to α-ketoglutaric acid results in formation of oxalacetic acid and glutamic acid. For alanine aminotransferase, the transfer of the amino group from alanine to α-ketoglutaric acid forms pyruvic acid and glutamic acid.

Although it was fi rst thought that alanine amino-transferase was found only in the cytoplasm of the hepatocyte, its activity is also found in mitochondria.1 Serum concentration of alanine aminotransferase is more specifi c to hepatocytes than is aspartate amino-transferase, which is found in other tissues such as skeletal muscle, heart muscle, small gut epithelial cells, the kidney, and erythrocytes. Moreover, isoforms of alanine aminotransferase, with diff erent distributions, have been reported.2

Alkaline phosphatase catalyses the hydrolysis of phosphate esters, and is found in biliary epithelium and the bile canalicular region of hepatocytes. Its function is not well established, but is thought to involve metabolite transport across cell membranes. Elevation of this enzyme can suggest intrahepatic or extrahepatic biliary obstruction, and occurs in hepatocellular injury to a lesser degree.

Gamma-glutamyltransferase transfers a gamma-glutamyl group between peptides, and is involved in aminoacid transfer. Although found in many tissues, its concentration is particularly high in the liver.

Isozymes of alkaline phosphatase, aspartate amino-transferase, and alanine aminotransferase can arise from other body tissues. Alkaline phosphatase is commonly raised in bony disease (eg, Paget’s disease or bony metastases). Isoenzyme studies can distinguish origin, but are rarely needed. Aspartate aminotransferase can be increased after muscle damage (eg, myocardial infarction). Alanine aminotransferase and gamma-glutamyltrans-ferase are not likely to be elevated for reasons other than liver dysfunction. Table 1 outlines possible causes of elevated liver enzymes and biochemical tests of function.

There are generally inconsistent relations between drug pharmacokinetics and serum biochemical liver function. Nevertheless, serum measurement of liver enzymes is inexpensive and gives useful information about liver injury, infl ammation, or cholestasis. Therefore, they are routinely used in clinical practice to make decisions about chemotherapy dose. Several studies have shown a relation between abnormal serum tests and toxic eff ects from chemotherapy, and are useful for current practice. These studies will be discussed in part II of this review.

Serum bilirubinSerum concentration of bilirubin is a marker of the liver’s ability to take up bilirubin from the plasma into the hepatocyte, conjugate it there with glucuronic acid, and excrete bilirubin glucuronides into bile. In adults with cancer, it is commonly a marker of cholestasis, when glucuronidated bilirubin is backed-up into plasma.

Bilirubin is the breakdown product of normal haem catabolism, most of which is derived from the haem of haemoglobin on splenic destruction of aged erythrocytes.

A

B

Figure 2: Histopathological injury secondary to chemotherapyMacrovesicular and microvesicular steatosis, ×400 magnifi cation (A, arrows) and sinusoidal dilatation, ×200 magnifi cation (B, arrows) in patients given neoadjuvant FOLFOX (fl uorouracil, leucovorin, and oxaliplatin) before liver resection for metastatic colorectal cancer. Cells in both panels stained with haematoxylin and eosin.

1094 www.thelancet.com/oncology Vol 9 November 2008

Review

Bilirubin, loosely bound to albumin in plasma to solubilise it, is taken up from the Disse spaces of liver sinusoids into hepatocytes, where it is esterifi ed at its propionyl sites with glucuronic acid under the catalytic activity of UGT1A1—a member of the family of uridine-diphosphoglucuronate transferase enzymes.

Conjugated or esterifi ed bilirubin is excreted into bile as water-soluble bilirubin diglucuronide. Elevated serum

concentrations of unconjugated bilirubin can occur in haemolysis (when rapid fl ux overwhelms removal capacity), or in hereditary defi ciencies of UGT activity (a common form of which is constitutional hyper-bilirubinaemia, or Gilbert’s syndrome, found in 5–7% of people of Caucasian origin). Elevated serum conjugated bilirubin implies regurgitation of bilirubin glucuronides from hepatocytes back into plasma, usually because of intrahepatic or extrahepatic obstruction to bile outfl ow and cholestasis. Raised levels are common in patients with cancer, and also occur as a result of drug-induced intrahepatic cholestatic injury.

The liver has substantial reserve capacity, and normal serum bilirubin levels can be maintained until there is enough injury to reduce the liver’s capacity to clear bilirubin from plasma. Serum concentration of bilirubin is very specifi c for potentially serious liver damage, and is an important indicator of liver functional loss.

Measures of liver synthetic function Albumin is the most abundant plasma protein, and is produced exclusively by the liver at a rate of 10–15 g a day. It transports small molecules and some drugs, and is important for maintenance of plasma oncotic pressure. The liver has large reserves of albumin synthetic capacity, and thus decreased serum albumin from liver impairment is highly important. Hypoalbuminaemia from liver-related causes is more common in chronic liver disease than in acute toxicity. Low serum albumin is commonly an adverse prognostic feature in many solid tumours.

Prothrombin time is a measure of the extrinsic pathway of the coagulation cascade. The international normalised ratio is the ratio of a patient’s prothrombin time to a normal (control) sample raised to a power, based on an international sensitivity index that corrects for variations between laboratories and reagents.

Prothrombin is one of several blood-coagulation proteins synthesised by the liver. The test is elevated most commonly when oral anticoagulants (ie, warfarin) are being given, or in vitamin K defi ciency. Increased prothrombin time because of liver dysfunction is a marker of severe synthetic abnormalities, which can be used to predict survival and outcomes in various disorders.3,4 These assessments in the setting of severe liver dysfunction will be discussed further in part 2 of our Review.

The so-called normal patient About 3% of healthy people have so-called abnormal aminotransferase levels because they are outside an SD of 2 compared with the normal Gaussian distribution for laboratory reference ranges.5 Age and sex can also aff ect upper reference limits for aminotransferase assessment.6 An analysis7 of more than 6000 healthy individuals found that, after exclusion of those who were hepatitis C positive or who drank excessive amounts of alcohol, 8% had had raised alanine aminotransferase or aspartate amino-transferase. Increasingly common reasons for this

Potential causes

Increased alkaline phosphatase

Bone disease (eg, Paget’s disease; hyperparathyroidism—any cause of increased bone turnover)Fracture repairThird trimester of pregnancy (ie, placental source)Increased parathyroid hormone (via eff ects on bone)Cirrhosis, especially primary biliary

Increased gamma-glutamyltransferase

AlcoholDrugs (eg, phenytoin, barbituates)Iron overloadFatty liver or obesityDiabetesMyocardial infarction

Increased alkaline phosphatase and gamma-glutamyltransferase

Marker of biliary-duct obstruction (multiple causes)Cholecystitis or cholelithiasisDrugs

Increased aspartate aminotransferase

Muscle-cell death (cardiac or skeletal)Haemolysis (aspartate aminotransferase present in erythrocytes)Hepatitis in alcoholism (rise in aspartate aminotransferase tends to be higher than that for alanine aminotransferase; ratio of former to latter of >2 suggests alcohol-related cause)Chronic liver disease or cirrhosis

Increased aspartate aminotransferase and alanine aminotransferase

Markers of hepatocellular damageHepatitis in alcoholismHepatitis (ie, viral, autoimmune, or drug-induced)Drugs (eg, isoniazid, statins, and amiodarone)Non-alcoholic steatohepatitisIschaemic or hypoxic liver damageAlanine aminotransferase more specifi c for liver damage than is aspartate aminotransferase; former is unlikely to be raised for non-hepatic causes

Increased bilirubin Prehepatic (ie, unconjugated hyperbilirubinaemia)• Haemolysis• Gilbert’s syndrome• Crigler-Najjar syndrome (types I and II)Intrahepatic (ie, conjugated hyperbilirubinaemia)• Drugs (eg, capecitabine and mitomycin)• Intrahepatic biliary obstructionPosthepatic (ie, conjugated hyperbilirubinaemia)• Biliary obstruction (any cause)

Decreased albumin Decreased production• Malnutrition• Malabsorption• Liver failure (any cause, usually chronic)• Infl ammatory statesIncreased loss• Nephrotic syndrome• Protein-losing enteropathy• BurnsRedistribution• Negative acute-phase protein (ie, serum level decreases with intercurrrent illness)• Ascites

Increased international normalised ratio or prothrombin time

Vitamin K defi ciencyWarfarin administrationCoagulopathy (eg, disseminated intravascular coagulation)Liver disease or cirrhosis (any cause)

Table 1: Common non-malignant causes of abnormal serum liver biochemical tests

www.thelancet.com/oncology Vol 9 November 2008 1095

Review

fi nding are presence of non-alcoholic fatty liver disease or the metabolic syndrome7,8 in people who are overweight or obese.

However, of greater concern is that conventional biochemical tests of liver function might fall within a defi ned normal range despite underlying histo-pathological liver damage. Reports9,10 have suggested that up to 16% of patients with chronic hepatitis C and 13% with non-alcoholic fatty liver disease might have normal amino transferase levels despite liver damage on histological analysis.

Furthermore, current standards for so-called normal alanine aminotransferase levels have been defi ned by use of populations that include individuals with subclinical liver disease such as hepatitis C and non-alcoholic fatty liver disease.8 Therefore, the current defi nition of normal alanine aminotransferase range might be inappropriate for use in all clinical situations, prompting suggestions that upper limits of normal should be lowered to identify more accurately those with subclinical liver disease,8 and for laboratories to be more careful during selection of population samples to represent a truly normal, healthy reference group.

Adverse-event criteria in oncology—are they clinically and pharmacologically relevant?The US National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) version 3.0 grades liver toxic eff ects on the basis of biochemical tests and some clinical measurements.11 The NCI’s compre-hensive, multimodality grading system for reporting of eff ects of cancer treatment has built on the strengths of previous versions.12 It is commonly and widely used in practice, particularly for clinical-trial assessment, screen-ing, and monitoring. The grades of toxic eff ects (1–5) are similar to other criteria such as those from WHO.

However, concerns have been raised about the clinical application and validation of some grades, especially for targeted therapies.13 Grade 3 adverse events are defi ned as “severe”; however, this category includes individuals with alanine aminotransferase, aspartate amino-transferase, alkaline phosphatase, or gamma-glutamyltransferase levels between fi ve-times and 20-times the upper limit of normal. To codify the severity scale of 1–5, the committee that defi nes common terminology criteria classifi es the above serum enzyme activity levels as grade 3 or “severe” adverse eff ects, and grade 4 as “life-threatening or disabling”.

However, patients with these fi ndings could be asymptomatic and levels might be reversible without serious organ damage, depending on disease course, the cause of elevations, and any associated clinical manifestations. So-called peak serum enzyme activity depends on how frequently measurements are made: at best, they are very poor predictors of severity of organ injury, future events, or drug eff ects; they are not measures of liver function.

Markers of liver injury must be distinguished from those of liver metabolism. Although both are highly important in practice and chemotherapy administration, they are measured by diff erent tests. Conventional liver testing is useful and important when screening for pre-existing liver injury or monitoring for development of drug-related toxic eff ects. However, serum enzyme activity is a less-useful prediction tool for assessment of adequate metabolic and excretory capacity in the liver during chemotherapy. Nevertheless, many clinicians use common toxicity criteria to predict the metabolic capacity of the liver, adjust chemotherapy doses, and avoid use of a particular drug. Safety is paramount in clinical practice, and caution is advised for drug use in patients with any liver injury or dysfunction because of the important role of this organ in drug pharmacokinetics.

Validated and cost-eff ective measures to assess liver metabolism are not yet routine in practice, and reliance on conventional liver tests has led to appropriate early-phase pharmacokinetic studies of drugs in the setting of biochemical abnormalities. However, metabolic and alternative tests of liver function that might predict more accurately an individual’s ability to metabolise a particular anticancer agent should be incorporated into adverse-event criteria in the future.

Causes of abnormal serum liver biochemistry in cancerDespite their shortcomings, standard biochemical tests of liver function are routinely used for chemotherapy dosing. For several reasons, they are commonly abnormal at the start of chemotherapy or during treatment. Possible causes of the abnormality must be considered in the interpretation of these tests: not all are directly associated with tumoral involvement of the liver.

Potential causes other than cancer or the chemotherapy agent should be considered when planning chemotherapy in a patient with abnormal liver biochemistry, or when a patient receiving chemotherapy develops abnormal serum liver tests. There are many potential causes of abnormal liver biochemical tests (table 1), and the work-up of a patient who presents initially with abnormal serum biochemistry will generally depend on whether the pattern of abnormality is cholestatic or hepatitic in nature.

Eff ect of cancer on liver function and biochemistryTumours in or around the liver and biliary tree can aff ect liver function, either by directly reducing the volume of functional healthy liver or by intrahepatic and extrahepatic biliary obstruction. Furthermore, portal-vein thrombosis from a hypercoagulable state or direct portal-vein infi ltration might occur, compromising vascular supply to healthy liver parenchyma and thus drug metabolism. A rare paraneoplastic disorder (called Stauff er’s syndrome) has been described with renal-cell carcinoma and is characterised by abnormal cholestatic biochemical tests, prolonged international normalised ratio, and

1096 www.thelancet.com/oncology Vol 9 November 2008

Review

hepatosplenomegaly without evidence of direct liver involvement by the tumour or of intrinsic liver disease.14

Humoral and immunological factors related to cancer probably contribute to cholestasis or infl ammatory sequelae in the liver. Histological evidence of liver damage has been shown in a small study15 of six autopsied patients with tumours that produced colony-stimulating-factors, interleukin 1, and interleukin 6.

Furthermore, cancer might indirectly aff ect drug metabolism in the liver. Cytochrome P4503A4 (CYP3A4) helps metabolise many chemotherapy agents and more than half of all drugs in clinical use.16 Cancer can generate a host infl ammatory response that involves cytokines such as interleukin 6 and tumour necrosis factor α, which manifests clinically by raised serum acute-phase reactants such as C-reactive protein, cachexia, and fever. This infl ammatory response is associated with decreased CYP3A4 activity in the liver, as measured in patients with cancer by use of an erythromycin breath test.17

In vitro evidence from transgenic mice suggests that reduced CYP3A4 activity occurs through transcriptional repression of the liver cytochrome P450 gene.18 The association between infl ammation and CYP3A4 activity has implications for the treatment of infl ammation in patients with cancer and its potential eff ect on chemotherapy drug metabolism. Furthermore, the infl ammatory response might aff ect the regulation and expression of drug transporters and thus drug clearance.19

Non-cancer drugs and social habitsMany prescription and non-prescription drugs can cause abnormal liver function before or during cancer treatment. The work-up of these patients will generally depend on whether the pattern of abnormality is cholestatic or hepatitic in nature. Complementary medicines can interact with anticancer agents: the most well known interaction is with St John’s wort—a potent CYP3A4 enzyme inducer.20 Importantly, up to 50% of patients with cancer might not tell their doctor about use of alternative medicines.21 If a patient has abnormal liver function, a careful history of medicine use and social habits must be taken.

Tobacco smoking can aff ect the functions of liver enzymes. Smoking signifi cantly lowers irinotecan exposure and treatment-induced neutropenia, potentially reducing drug eff ectiveness most likely by induction of drug-metabolising enzymes;22 smoking also increases erlotinib clearance.23 Alcohol is one of the most common causes of abnormal serum liver biochemistry. Consumption commonly increases aspartate amino transferase (more so than alanine aminotransferase; a ratio of aspartate aminotransferase to alanine aminotransferase of more than 2 increases the likelihood that alcohol caused the abnormality) and raises gamma-glutamyltransferase. Accurate quantifi cation of alcohol intake is important in the overall assessment of risk and in identifi cation of the causes of biochemical abnormalities.

Patient demographicsObesity is increasing worldwide and oncologists must be aware of the potential for subclinical or overt liver dysfunction in these individuals. Non-alcoholic fatty liver disease is common, aff ecting 25–35% of the general population and up to 80% of patients with obesity. About 10% of people with this disease have an active form called non-alcoholic steatohepatitis that can progress to cirrhosis or hepatocellular carcinoma, or can necessitate liver transplantation.24 The metabolic syndrome is related to obesity and with non-alcoholic fatty liver disease.

Ageing aff ects liver function: liver mass, blood fl ow, and activity of P450 enzymes decrease with age.25,26 Albumin decreases by about 15% with age,26 which might aff ect the amount of free or unbound drug in plasma and the volume of distribution; however, the magnitude of this eff ect is thought to be minor. Toxic eff ects of chemotherapy can be high in elderly patients, particularly those who are frail, partly due to multiple comorbidities.25 These individuals might take many medicines, increasing the potential for drug interactions.

Viral hepatitis Viral hepatitis can contribute to morbidity and mortality in patients who are receiving chemotherapy. Hepatitis B reactivation can occur during chemotherapy and treatment with high-dose steroids, which in severe cases can lead to fulminant hepatic necrosis and death.27 Lamivudine prophylaxis signifi cantly reduces reactivation frequency28 and should be started before chemotherapy in those at risk. Hepatitis B serology should be done for any patient who might be at risk of viral exposure, and for every patient undergoing intensive chemotherapy.

Compared with hepatitis B infection, the frequency of overt liver dysfunction in patients with hepatitis C who are receiving chemotherapy is very uncommon. No prophylaxis is established, although treatment of chronic hepatitis C with interferon alpha and ribavirin has been eff ective in recipients of allogeneic bone-marrow transplantation.29

Chemotherapy-induced hepatotoxicityMany drugs can cause transient changes in liver-biochemical tests. Most adverse reactions are idiosyncratic and are due to individual patient diff erences in susceptibility to drug-induced liver injury or inability to recover from the injury. Most patients tolerate the drug or agent, or can adapt to it. However, some drugs in particular cause specifi c toxic eff ects in the liver.

A serious liver-related complication of high-dose chemotherapy, most commonly in those who have had allogeneic bone-marrow transplantation, is hepatic veno-occlusive disease or sinusoidal obstruction syndrome. This disease is characterised by jaundice and hyper-bilirubinaemia, painful hepatomegaly, and weight gain from fl uid retention. Histological analysis shows obliteration of the venular lumina by loose immature fi brous tissue intermixed with haemorrhage and fi brin

www.thelancet.com/oncology Vol 9 November 2008 1097

Review

(fi gure 3), which can be accompanied by surrounding hepatocellular congestion, atrophy, necrosis, and (ultimately) fi brosis.

More than 30% of patients with this disease die,30 rising to 90% in severe cases.31 Frequency of veno-occlusive disease after allogeneic transplantation varies between series, but is around 10–25%.30 It can also occur after autologous stem-cell transplantation, but with lower frequency. Pre-existing liver disease or previous liver irradiation might increase susceptibility.32 Preventive strategies such as careful selection of patients for transplant, and the use of heparin and ursodiol (ursodeoxycholic acid, a bile salt) are under investigation. Although treatment is mainly supportive, encouraging results have been noted for defi brotide—an anti-thrombotic, anti-infl ammatory drug that modulates platelet activity, promotes fi brinolysis, and prevents fi brin deposition.31 Table 233–70 outlines chemotherapy drugs that have been associated with liver dysfunction; some commonly used anticancer agents are discussed below.

Methotrexate A folate antimetabolite, methotrexate is widely known to be associated with liver dysfunction. Therefore, serum liver biochemistry should be monitored regularly during this treatment. It is mainly excreted unchanged in urine. A small proportion is hydroxylated or metabolised to poly-glutamate derivatives; liver cells seem to retain some of the methotrexate polyglutamates for prolonged periods.65

Methotrexate can induce various types of hepatotoxicity, including liver atrophy, necrosis, cirrhosis, fatty changes, and periportal fi brosis. One study63 reported a 14% frequency of increased aminotransferases after metho-trexate treatment for gestational trophoblastic disease. Increased aminotransferases are more likely with transient high doses, whereas chronic low doses (eg, for rheumatoid arthritis or immunosuppression) are associated with fi brosis or cirrhosis.64 Methotrexate-induced hepatotoxicity seems more likely in those with increased body-mass index.66 A study71 of mice showed that methotrexate-induced liver fi brosis is mediated through an adenosine pathway, and via interaction between adenosine (an endogenous regulator of infl ammation and cell repair) and hepatocyte A2 receptors. In the presence of ascites, the volume of distribution of this drug changes substantially, increasing the likelihood of toxic eff ects.72

Oxaliplatin A platinum analogue that is widely used in adjuvant and advanced settings in colorectal cancer, oxaliplatin is mainly excreted through the kidneys and does not undergo cytochrome P450 or liver-mediated metabolism. It is tolerated well by patients who have liver dysfunction.67 Nevertheless, studies59,60,68,73 of patients undergoing liver metastectomy after neoadjuvant oxaliplatin-based treatment have shown histological evidence of liver damage (fi gure 2).

Neoadjuvant chemotherapy with fl uorouracil and oxaliplatin-based regimens has been associated with vascular changes in healthy liver parenchyma, more so than steatosis.73 Frequency of oxaliplatin-associated sinusoidal obstruction or dilatation syndrome varies in reports from 20% to 80%.59,60,68 This condition might increase morbidity, but has not been shown to increase mortality after liver metastectomy. However, current recommendations are to limit the number of preoperative cycles if possible.

Irinotecan A topoisomerase I inhibitor, irinotecan is commonly given with fl uorouracil for metastatic colorectal cancer. It is inactivated by glucuronidation catalysed by UGT1A1. Toxic levels of irinotecan can accumulate from standard doses if glucuronidation is reduced, as occurs in Gilbert’s

A

B

Figure 3: Hepatic veno-occlusive disease after chemotherapy and abdominal radiotherapy(A) Fibrosis and near-obliteration of venule (arrow), with atrophy and congestion of hepatic parenchyma (haematoxylin and eosin stain, ×200 magnifi cation). (B) Collagen of venule wall (arrow) and loose fi brosis has almost obliterated venule lumen (Masson trichrome stain, ×200 magnifi cation).

1098 www.thelancet.com/oncology Vol 9 November 2008

Review

Liver toxic eff ects Frequency Severity References

Asparaginase Inhibition of protein synthesis (decreased albumin and increased aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, and bilirubin)

Common Mostly reversible 33

Steatosis Very common (up to 87% of patients) Mostly reversible

Azathioprine and mercaptopurine

Cholestatic jaundice or intrahepatic cholestasis Increased aminotransferases

Common (more so with mercaptopurine than with azathioprine)

Mostly reversible 34

Busulfan Veno-occlusive disease in high doses or in transplantation Common in bone-marrow transplantation (10–25% of patients)

Severe and life-threatening 30,35

Cholestatic hepatitis Rare case reports Reversible 35

Capecitabine Hyperbilirubinaemia, usually without increased alkaline phosphatase and gamma-glutamyltransferase; might be related to haemolysis

Common (23–25% of patients) Grade 3–4 in up to 23% of patients

36–38

Carmustine Increased aminotransferases and alkaline phosphatase, with or without increased bilirubin

Common, up to 26% of patients Usually mild and reversible; rare fatalities

39

Chlorambucil Fibrosis and cirrhosis Rare, anecdotal Case reports of severe damage (but viral infection not excluded)

40

Cisplatin Increased aminotransferases Common (more so with high doses) Usually transient 41

Steatosis and cholestasis Rare Usually transient 41

Cyclophosphamide Veno-occlusive disease in high doses or in transplantation Common in bone-marrow transplantation (10–25% of patients)

Severe and life-threatening 30,42

Isolated case reports of idiosyncratic reactions Uncommon Usually transient 42

Cytarabine Increased aspartate aminotransferase and alanine aminotransferase; cholestatic jaundice and intrahepatic cholestasis

Common Usually reversible 43

Dacarbazine Case reports of fulminant liver failure (thrombotic occlusions) Rare Can be life-threatening 44

Dactinomycin Increased aminotransferases in children Up to 17% of patients Can be life-threatening 45

Doxorubicin Idiosyncratic reactions, including increased aminotransferases and bilirubin

Rare Usually transient 46

Etoposide Veno-occlusive disease in high doses or in transplantation Common in bone-marrow transplantation (10–25% of patients)

Severe and life-threatening 30,47

Case reports of severe hepatocellular damage at standard doses Rare Severe 47

Fluorouracil Steatosis Common Usually subclinical 48,49

Hepatotoxicity Rare Usually subclinical 48,49

Floxuridine (into hepatic artery)

Increased aminotransferases, alkaline phosphatase, and bilirubin Common, transaminitis in up to 50% of patients

Hepatitis improves after cessation

50,51

Biliary stricture or sclerosis Up to 16% of patients Secondary sclerosing cholangitis irreversible

51

Gefi tinib Increased aminotransferases Uncommon Usually subclinical 52

Gemcitabine Increased aminotransferases Very common, up to 60% of patients Generally transient and reversible 53,54

Case reports of fatal cholestatic hepatotoxicity Rare Can be fatal 53,54

Imatinib Increased aminotransferases or bilirubin Up to 10% of patients 2–6% grade 4 55,56

Liver necrosis or failure Rare Case reports of fatalities 55,56

Interferon Increased aminotransferases Common Usually mild and improve after cessation

57

Interleukin 2 Increased bilirubin (ie, intrahepatic cholestasis) Common Usually reversible 58

Increased aminotransferases and alkaline phosphatase Common Usually reversible 58

Irinotecan Steatosis and steatohepatitis Common (25–50% of patients) Steatohepatitis can increase morbidity if used before liver resection

48,59–61

Increased aminotransferases and bilirubin Up to 25% of patients Usually reversible 48,59–61

Melphalan Hepatotoxicity, thrombus, and veno-occlusive disease with isolated liver perfusion

Rare, but veno-occlusive disease common in bone-marrow transplantation (10–25% of patients)

Veno-occlusive disease severe and life-threatening

30,62

Transient enzyme increase Common at high doses Usually transient 62

Methotrexate Increased aspartate aminotransferase and alanine aminotransferase at high doses

Common with high dose Transient and reversible 63–66

Liver atrophy, necrosis, cirrhosis, fatty changes, and periportal fi brosis at chronic low dose

More common with chronic use (eg, in rheumatoid arthritis) and cumulative dose >2 g

Potentially irreversible 63–66

(Continues on next page)

www.thelancet.com/oncology Vol 9 November 2008 1099

Review

syndrome. Irinotecan pharmacokinetics are complex, but well studied.

Its use is associated with hepatic steatosis and steatohepatitis in up to 50% of liver-resection samples.60,61 Steatohepatitis, the more-severe condition, has been associated with increased morbidity, and increased mortality in some series, after liver resection. A review59 of 406 patients undergoing neoadjuvant chemotherapy before resection of colorectal liver metastases noted that signifi cantly more patients with steatohepatitis died after 90 days mortality compared with those without the disease (15% vs 2%, p=0·001).

It is generally more common to use oxaliplatin-based regimens than those that are irinotecan-based for neoadjuvant treatment before liver resection.

Irinotecan can also cause mild elevations of serum transaminases and bilirubin in up to 25% of patients.74

Fluorouracil A uracil analogue and antimetabolite that is metabolised mainly in the liver, fl uorouracil is commonly given with oxaliplatin and irinotecan for colorectal cancer. Fluorouracil has been associated with hepatic steatosis on radiological and histological fi ndings,48,49 but has not been shown to aff ect morbidity or mortality by contrast with steatohepatitis.75

Capecitabine is an oral prodrug that is converted to fl uorouracil by thymidine phosphorylase, an enzyme present in higher levels in some tumours and the liver than in other healthy tissue. Hyperbilirubinaemia is common and usually reversible: in clinical trials of early-stage and metastatic colon cancer, grade 3 hyper-bilirubinaemia was present in 18–23% of patients and grade 4 in 1–5%.37,38 This eff ect is not usually accompanied by rises to the same degree in transaminases, and is not thought to be associated with hepatobiliary dysfunction; rather, it might be related to haemolysis.36

Gemcitabine A pyrimidine analogue that is used to treat various cancers such as non-small-cell lung cancer and pancreatic cancer, gemcitabine is associated with a mild rise in transaminases in up to 60% of patients, and in serum bilirubin in a small proportion of patients. These changes are usually transient, but are very rarely associated with severe hepatotoxicity.53,54 Monitoring of liver tests in patients who

are receiving gemcitabine is recommended; expected mild transaminitis is usually of no clinical consequence.

Imatinib An oral small-molecule tyrosine kinase inhibitor specifi cally developed to inhibit the BCR-ABL kinase in chronic myeloid leukaemia, imatinib also eff ectively inhibits the KIT tyrosine kinase and platelet-derived growth factor receptor in gastrointestinal stromal tumours. Cases of severe liver injury, including liver failure and necrosis, have been recorded for imatinib, and regular monitoring of serum liver tests is important.

Medical history• Age• Family history of drug toxic effects• Personal history of viral hepatitis• Risk factors for viral hepatitis

Social habit history• Alcohol• Recreational drugs• Tobacco

Medical history • Prescription medicines• Over-the-counter medicines• Complementary and alternative medicines• Previous allergies or toxic effects

Examination• Body-mass index• Gastrointestinal, skin, and whole body-surface examination for signs of chronic liver disease

Basic investigations• Serum liver biochemical tests at baseline• Repeat if abnormal• Further directed investigation depends on pattern of abnormality (ie, cholestatic or hepatitic)• Consider hepatology consultation if abnormalities unexplained

Consider hepatitis serology• Risk factors• Planning high-dose chemotherapy

Imaging• If appropriate for tumour staging• Further directed imaging if sustained biochemical abnormalities on liver function tests—eg, ultrasonography or magnetic resonance cholangiopancreatography for obstruction

Chemotherapy choices• Awareness of drugs that might cause hepatotoxicity (table 2) and select appropriately• Frequent monitoring of liver tests with immediate (ie, <48 h) repeat if normal• Limit number of preoperative cycles before planned liver resection (eg, oxaliplatin for colorectal cancer)• Use currently available dosing strategies if biochemical liver-test abnormalities exist

Figure 4: Assessment before chemotherapy: identifi cation of patients at risk of hepatotoxicity

Liver toxic eff ects Frequency Severity References

(Continued from previous page)

Oxaliplatin Vascular changes; sinusoidal obstruction or dilatation syndrome Common (20–80% of patients) Might increase morbidity, but not mortality, after liver resection

59,60,67,68

Paclitaxel Increased bilirubin, aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase

For high doses, up to 37% of patients Generally reversible 69

Topotecan Increased aminotransferases and alkaline phosphatase Up to 5–8% of patients Low-grade and reversible 70

Table 2: Anticancer agents associated with toxic liver eff ects

1100 www.thelancet.com/oncology Vol 9 November 2008

Review

Test abnormalities occur in up to 10% of patients, 2–6% of whom have grade 3–4 toxic eff ects.55,56

Conclusion: inadequacies in measurement of liver functionLiver metabolism plays a very important part in the pharmacokinetics of many chemotherapy agents—especially their activation, degradation, and excretion. Conventional techniques for assessment of liver function do not measure metabolic activity of the liver—a cause for concern. So-called normal serum liver biochemistry might not identify patients with impaired liver function, and conventional chemotherapy dosing in these patients could be deleterious or ineff ective.

We have focused on the importance and meaning of serum biochemical liver tests, many of which are not true measures of liver function. We have also discussed the causes of abnormal biochemical tests in patients with cancer—both chemotherapy-related and those not associated with this treatment. Figure 4 outlines the steps that should be taken before chemotherapy to routinely consider the risk of hepatotoxicity and the potential eff ects of altered liver metabolism in patients.

Chemotherapy agents might be directly hepatotoxic, and cancer has direct and indirect eff ects on liver function; these factors are commonly not considered as part of dosing recommendations. Moreover, dosing based on body surface area has drawbacks and does not account for variability between patients of drug metabolism or clearance. Alternative methods for assessment of liver function need investigation for optimisation of dosing for cytotoxics. Individualised dosing on the basis of a patient’s drug metabolic, excretory, and pharmacodynamic genotype and phenotype could be a logical path to this goal.

Confl icts of interestThe authors declared no confl icts of interest.

References1 Swick RW, Barnstein PL, Stange JL. The metabolism of

mitochondrial proteins. I. Distribution and characterization of the isozymes of alanine aminotransferase in rat liver. J Biol Chem 1965; 240: 3334–41.

2 Lindblom P, Rafter I, Copley C, et al. Isoforms of alanine aminotransferases in human tissues and serum—diff erential tissue expression using novel antibodies. Arch Biochem Biophys 2007; 466: 66–77.

Search strategy and selection criteria

We searched PubMed and included original articles and published reviews with their references. The specifi ed search terms were: “liver and chemotherapy”; “hepatotoxicity and chemotherapy”; “liver functional imaging”; and “x (chemotherapy drug) and hepatotoxicity”, for relevant papers published up to September, 2008. Only papers published in English were selected. Abstracts or meeting reports were included only if they were directly relevant to the discussion and if the published data were unavailable.

3 Garcia-Pagan JC, Heydtmann M, Raff a S, et al. TIPS for Budd-Chiari syndrome: long-term results and prognostics factors in 124 patients. Gastroenterology 2008; published online May 21. DOI:10.1053/j.gastro.2008.05.051.

4 Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001; 33: 464–70.

5 Giannini EG, Testa R, Savarino V. Liver enzyme alteration: a guide for clinicians. CMAJ 2005; 172: 367–79.

6 Siest G, Henry J, Schiele F, Young D. Interpretation of clinical laboratory tests: reference values and their biological variation. Foster City: Biomedical Publications, 1985.

7 Ioannou GN, Boyko EJ, Lee SP. The prevalence and predictors of elevated serum aminotransferase activity in the United States in 1999–2002. Am J Gastroenterol 2006; 101: 76–82.

8 Prati D, Taioli E, Zanella A, et al. Updated defi nitions of healthy ranges for serum alanine aminotransferase levels. Ann Intern Med 2002; 137: 1–10.

9 Gholson CF, Morgan K, Catinis G, et al. Chronic hepatitis C with normal aminotransferase levels: a clinical histologic study. Am J Gastroenterol 1997; 92: 1788–92.

10 Mofrad P, Contos MJ, Haque M, et al. Clinical and histologic spectrum of nonalcoholic fatty liver disease associated with normal ALT values. Hepatology 2003; 37: 1286–92.

11 US National Cancer Institute. Common terminology criteria for adverse events version 3.0 (CTCAE). Aug 9, 2006. http://ctep.cancer.gov/forms/CTAEv3.pdf (accessed Sept 4, 2008).

12 Trotti A, Colevas AD, Setser A, et al. CTCAE v3.0: development of a comprehensive grading system for the adverse eff ects of cancer treatment. Semin Radiat Oncol 2003; 13: 176–81.

13 Edgerly M, Fojo T. Is there room for improvement in adverse event reporting in the era of targeted therapies? J Natl Cancer Inst 2008; 100: 240–42.

14 Giannakos G, Papanicolaou X, Trafalis D, Michaelidis I, Margaritis G, Christofi lakis C. Stauff er’s syndrome variant associated with renal cell carcinoma. Int J Urol 2005; 12: 757–59.

15 Suzuki A, Takahashi T, Okuno Y, et al. Liver damage in patients with colony-stimulating factor-producing tumors. Am J Med 1993; 94: 125–32.

16 Bertz RJ, Granneman GR. Use of in vitro and in vivo data to estimate the likelihood of metabolic pharmacokinetic interactions. Clin Pharmacokinet 1997; 32: 210–58.

17 Rivory LP, Slaviero KA, Clarke SJ. Hepatic cytochrome P450 3A drug metabolism is reduced in cancer patients who have an acute-phase response. Br J Cancer 2002; 87: 277–80.

18 Charles KA, Rivory LP, Brown SL, Liddle C, Clarke SJ, Robertson GR. Transcriptional repression of hepatic cytochrome P450 3A4 gene in the presence of cancer. Clin Cancer Res 2006; 12: 7492–97.

19 Petrovic V, Teng S, Piquette-Miller M. Regulation of drug transporters during infection and infl ammation. Mol Interv 2007; 7: 99–111.

20 Markowitz JS, Donovan JL, DeVane CL, et al. Eff ect of St John’s wort on drug metabolism by induction of cytochrome P450 3A4 enzyme. JAMA 2003; 290: 1500–04.

21 MacLennan AH, Myers SP, Taylor AW. The continuing use of complementary and alternative medicine in South Australia: costs and beliefs in 2004. Med J Aust 2006; 184: 27–31.

22 de Jong FA, van der Bol JM, Mathijssen RH, et al. Cigarette smoking during irinotecan therapy: eff ects on pharmacokinetics and neutropenia. Proc Am Soc Clin Oncol 2007; 25 (suppl): abstr 2506.

23 Hamilton M, Wolf JL, Rusk J, et al. Eff ects of smoking on the pharmacokinetics of erlotinib. Clin Cancer Res 2006; 12: 2166–71.

24 Preiss D, Sattar N. Non-alcoholic fatty liver disease: an overview of prevalence, diagnosis, pathogenesis and treatment considerations. Clin Sci 2008; 115: 141–50.

25 Balducci L, Extermann M. Management of cancer in the older person: a practical approach. Oncologist 2000; 5: 224–37.

26 Hunter CP Johnson KA, Muss HB (eds). Cancer in the elderly. New York/Basel: Informa Healthcare, 2000.

27 Sarrecchia C, Cappelli A, Aiello P. HBV reactivation with fatal fulminating hepatitis during rituximab treatment in a subject negative for HBsAg and positive for HBsAb and HBcAb. J Infect Chemother 2005; 11: 189–91.

www.thelancet.com/oncology Vol 9 November 2008 1101

Review

28 Yeo W, Ho WM, Hui P, et al. Use of lamivudine to prevent hepatitis B virus reactivation during chemotherapy in breast cancer patients. Breast Cancer Res Treat 2004; 88: 209–15.

29 Peff ault de Latour R, Asselah T, Levy V, et al. Treatment of chronic hepatitis C virus in allogeneic bone marrow transplant recipients. Bone Marrow Transplant 2005; 36: 709–13.

30 Carreras E. Veno-occlusive disease of the liver after hemopoietic cell transplantation. Eur J Haematol 2000; 64: 281–91.

31 Ho VT, Revta C, Richardson PG. Hepatic veno-occlusive disease after hematopoietic stem cell transplantation: update on defi brotide and other current investigational therapies. Bone Marrow Transplant 2008; 41: 229–37.

32 Laurie R, Chau M, Dow C. Radiation induced liver disease: is hereditary haemochromatosis a risk factor? J Radiother Prac 2003; 3: 101–04.

33 Chabner B, Myers CE. Antitumour antibiotics. 4th edn. Philadelphia: JB Lippincott, 1993.

34 Einhorn M, Davidsohn I. Hepatotoxicity of mercaptopurine. JAMA 1964; 188: 802–06.

35 King PD, Perry MC. Hepatotoxicity of chemotherapeutic and oncologic agents. Gastroenterol Clin North Am 1995; 24: 969–90.

36 Nikolic-Tomasevic Z, Jelic S, Cassidy J, Filipovic-Ljeskovic I, Tomasevic Z. Fluoropyrimidine therapy: hyperbilirubinemia as a consequence of hemolysis. Cancer Chemother Pharmacol 2005; 56: 594–602.

37 Scheithauer W, McKendrick J, Begbie S, et al. Oral capecitabine as an alternative to iv 5-fl uorouracil-based adjuvant therapy for colon cancer: safety results of a randomized, phase III trial. Ann Oncol 2003; 14: 1735–43.

38 Van Cutsem E, Twelves C, Cassidy J, et al. Oral capecitabine compared with intravenous fl uorouracil plus leucovorin in patients with metastatic colorectal cancer: results of a large phase III study. J Clin Oncol 2001; 19: 4097–106.

39 De Vita VT, Carbone PP, Owens AH Jr, Gold GL, Krant MJ, Edmonson J. Clinical trials with 1,3-bis(2-chloroethyl)-1-nitrosourea, NSC-409962. Cancer Res 1965; 25: 1876–81.

40 Koler RD, Forsgren AL. Hepatotoxicity due to chlorambucil: report of a case. JAMA 1958; 167: 316–17.

41 Hill JM, Loeb E, MacLellan A, Hill NO, Khan A, King JJ. Clinical studies of platinum coordination compounds in the treatment of various malignant diseases. Cancer Chemother Rep 1975; 59: 647–59.

42 Juma FD. Eff ect of liver failure on the pharmacokinetics of cyclophosphamide. Eur J Clin Pharmacol 1984; 26: 591–93.

43 Slavin RE, Dias MA, Saral R. Cytosine arabinoside induced gastrointestinal toxic alterations in sequential chemotherapeutic protocols: a clinical-pathologic study of 33 patients. Cancer 1978; 42: 1747–59.

44 Asbury RF, Rosenthal SN, Descalzi ME, Ratcliff e RL, Arseneau JC. Hepatic veno-occlusive disease due to DTIC. Cancer 1980; 45: 2670–74.

45 Pritchard J, Raine J, Wallendszus K. Hepatotoxicity of actinomycin-D. Lancet 1989; 1: 168.

46 Aviles A, Herrera J, Ramos E, Ambriz R, Aguirre J, Pizzuto J. Hepatic injury during doxorubicin therapy. Arch Pathol Lab Med 1984; 108: 912–13.

47 Tran A, Housset C, Boboc B, Tourani JM, Carnot F, Berthelot P. Etoposide (VP 16-213) induced hepatitis. Report of three cases following standard-dose treatments. J Hepatol 1991; 12: 36–39.

48 Moertel CG, Fleming TR, Macdonald JS, Haller DG, Laurie JA. Hepatic toxicity associated with fl uorouracil plus levamisole adjuvant therapy. J Clin Oncol 1993; 11: 2386–90.

49 Peppercorn PD, Reznek RH, Wilson P, Slevin ML, Gupta RK. Demonstration of hepatic steatosis by computerized tomography in patients receiving 5-fl uorouracil-based therapy for advanced colorectal cancer. Br J Cancer 1998; 77: 2008–11.

50 Doria MI Jr, Shepard KV, Levin B, Riddell RH. Liver pathology following hepatic arterial infusion chemotherapy. Hepatic toxicity with FUDR. Cancer 1986; 58: 855–61.

51 Hohn D, Melnick J, Stagg R, et al. Biliary sclerosis in patients receiving hepatic arterial infusions of fl oxuridine. J Clin Oncol 1985; 3: 98–102.

52 Ho C, Davis J, Anderson F, Bebb G, Murray N. Side eff ects related to cancer treatment: case 1. Hepatitis following treatment with gefi tinib. J Clin Oncol 2005; 23: 8531–33.

53 Robinson K, Lambiase L, Li J, Monteiro C, Schiff M. Fatal cholestatic liver failure associated with gemcitabine therapy. Dig Dis Sci 2003; 48: 1804–08.

54 Saif MW, Shahrokni A, Cornfeld D. Gemcitabine-induced liver fi brosis in a patient with pancreatic cancer. JOP 2007; 8: 460–67.

55 Ridruejo E, Cacchione R, Villamil AG, Marciano S, Gadano AC, Mando OG. Imatinib-induced fatal acute liver failure. World J Gastroenterol 2007; 13: 6608–11.

56 Mindikoglu AL, Regev A, Bejarano PA, Martinez EJ, Jeff ers LJ, Schiff ER. Imatinib mesylate (gleevec) hepatotoxicity. Dig Dis Sci 2007; 52: 598–601.

57 Quesada JR, Talpaz M, Rios A, Kurzrock R, Gutterman JU. Clinical toxicity of interferons in cancer patients: a review. J Clin Oncol 1986; 4: 234–43.

58 Fisher B, Keenan AM, Garra BS, et al. Interleukin-2 induces profound reversible cholestasis: a detailed analysis in treated cancer patients. J Clin Oncol 1989; 7: 1852–62.

59 Vauthey JN, Pawlik TM, Ribero D, et al. Chemotherapy regimen predicts steatohepatitis and an increase in 90-day mortality after surgery for hepatic colorectal metastases. J Clin Oncol 2006; 24: 2065–72.

60 Morris-Stiff G, Tan YM, Vauthey JN. Hepatic complications following preoperative chemotherapy with oxaliplatin or irinotecan for hepatic colorectal metastases. Eur J Surg Oncol 2008; 34: 609–14.

61 Fernandez FG, Ritter J, Goodwin JW, Linehan DC, Hawkins WG, Strasberg SM. Eff ect of steatohepatitis associated with irinotecan or oxaliplatin pretreatment on resectability of hepatic colorectal metastases. J Am Coll Surg 2005; 200: 845–53.

62 Lazarus HM, Herzig RH, Graham-Pole J, et al. Intensive melphalan chemotherapy and cryopreserved autologous bone marrow transplantation for the treatment of refractory cancer. J Clin Oncol 1983; 1: 359–67.

63 Berkowitz RS, Goldstein DP, Bernstein MR. Ten year’s experience with methotrexate and folinic acid as primary therapy for gestational trophoblastic disease. Gynecol Oncol 1986; 23: 111–18.

64 King PD, Perry MC. Hepatotoxicity of chemotherapy. Oncologist 2001; 6: 162–76.

65 Tian H, Cronstein BN. Understanding the mechanisms of action of methotrexate: implications for the treatment of rheumatoid arthritis. Bull NYU Hosp Jt Dis 2007; 65: 168–73.

66 Hoekstra M, van Ede AE, Haagsma CJ, et al. Factors associated with toxicity, fi nal dose, and effi cacy of methotrexate in patients with rheumatoid arthritis. Ann Rheum Dis 2003; 62: 423–26.

67 Doroshow JH, Synold TW, Gandara D, et al. Pharmacology of oxaliplatin in solid tumor patients with hepatic dysfunction: a preliminary report of the National Cancer Institute Organ Dysfunction Working Group. Semin Oncol 2003; 30 (suppl 15): 14–19.

68 Rubbia-Brandt L, Audard V, Sartoretti P, et al. Severe hepatic sinusoidal obstruction associated with oxaliplatin-based chemotherapy in patients with metastatic colorectal cancer. Ann Oncol 2004; 15: 460–66.

69 Huizing MT, Misser VH, Pieters RC, et al. Taxanes: a new class of antitumor agents. Cancer Invest 1995; 13: 381–404.

70 Creemers GJ, Lund B, Verweij J. Topoisomerase I inhibitors: topotecan and irenotecan. Cancer Treat Rev 1994; 20: 73–96.

71 Chan ES, Montesinos MC, Fernandez P, et al. Adenosine A(2A) receptors play a role in the pathogenesis of hepatic cirrhosis. Br J Pharmacol 2006; 148: 1144–55.

72 Li J, Gwilt P. The eff ect of malignant eff usions on methotrexate disposition. Cancer Chemother Pharmacol 2002; 50: 373–82.

73 Aloia T, Sebagh M, Plasse M, et al. Liver histology and surgical outcomes after preoperative chemotherapy with fl uorouracil plus oxaliplatin in colorectal cancer liver metastases. J Clin Oncol 2006; 24: 4983–90.

74 Burris HA 3rd, Fields SM. Topoisomerase I inhibitors. An overview of the camptothecin analogs. Hematol Oncol Clin North Am 1994; 8: 333–55.

75 Zorzi D, Laurent A, Pawlik TM, Lauwers GY, Vauthey JN, Abdalla EK. Chemotherapy-associated hepatotoxicity and surgery for colorectal liver metastases. Br J Surg 2007; 94: 274–86.