REVIEW CME EDUCATIONAL OBJECTIVE: Readers will consider when further evaluation of mildly abnormal liver enzyme CREDIT levels is warranted George Aragon, MD Zobair M. Younossi, MD, MPH Center for Liver Diseases, Inova Fairfax Center for Liver Diseases, Inova Fairfax Hospital, Hospital, Falls Church, VA; Department Falls Church, VA of Gastroenterology, George Washington University, Washington, DC

When and how to evaluate mildly elevated liver enzymes in apparently healthy patients

■■ABSTRACT n seemingly healthy patients, abnormal Iliver enzyme levels challenge even the most Because 1% to 9% of people without symptoms have experienced clinicians in deciding what further elevated liver enzymes, extensive evaluation of all abnor- evaluation to pursue, if any. Automated labora- mal test results would expose many patients to undue tory testing has made serum liver enzyme levels risks and expenses. On the other hand, failure to evalu- very easy to obtain, leading to an increase in ate minor liver enzyme elevations could mean missing testing and also in the number of incidental the early diagnosis of potentially treatable disorders. This abnormal findings. An estimated 1% to 9% of review discusses likely causes of elevated aminotransfer- people who have no symptoms have high liver enzyme levels when screened with standard ase, alkaline phosphatase, and gamma-glutamyl trans- 1,2 2 ferase levels and provides algorithms for evaluating high biochemistry panels. A US survey showed elevated alanine aminotransferase (ALT) in liver enzyme values in apparently healthy patients in the 8.9% of surveyed people from 1999 to 2002, an primary care setting. increase from previous reports. ■■KEY POINTS An extensive evaluation can be costly, anx- iety-provoking, and risky, especially if it leads Nonalcoholic fatty liver disease is the most common to unnecessary invasive procedures such as liver cause of asymptomatic elevated aminotransferase levels. biopsy or endoscopic retrograde cholangiopan- creatography (ERCP). Not all people with a single, isolated, mildly elevated liver enzyme Suspect alcoholic liver disease when the aminotransfer- value have underlying liver disease, nor do ases are elevated and the aspartate aminotransferase they require an extensive evaluation. Factors level is two to three times higher than the alanine ami- to consider when deciding whether to evaluate notransferase level, especially when gamma-glutamyl include: transferase levels are elevated. • The patient’s overall health, including chronic illness If medications or alcohol is a suspected cause of elevat- • The duration and pattern of enzyme eleva- ed aminotransferase levels, remeasure the levels after 6 tion to 8 weeks of abstinence. • Patient characteristics such as age, a per- sonal or family history of liver, lung, or neurologic disease, risk factors for viral hepatitis, amount of alcohol consumption, use of prescribed or over-the-counter drugs or dietary supplements • The costs and risks associated with addi- doi:10.3949/ccjm.77a.09064 tional evaluation.

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 195 mildly Elevated liver enzymes

ases typically indicates hepatocellular injury, Table 1 whereas elevated alkaline phosphatase and Liver diseases and associated GGT indicates cholestatic injury. Elevated al- liver enzyme elevations kaline phosphatase and aminotransferases can indicate a mixed pattern of injury. Hepatocellular diseases (aminotransferase elevations predominate) High AST, ALT suggest liver cell damage Common Both aspartate aminotransferase (AST) and Nonalcoholic fatty liver disease ALT are normally present in serum at low lev- Chronic viral hepatitis els, usually less than 30 to 40 U/L. Although Genetic hemochromatosis the actual values may differ from laboratory to (northern European ethnicity) Alcoholic liver disease laboratory, normal serum levels are usually less than 40 U/L for AST and less than 50 U/L for Medication toxicity (see Table 2) Autoimmune hepatitis ALT. On the other hand, some experts have Less common suggested lowering the upper limit of normal Wilson disease because of the increasing rate of obesity and Alpha-1-antitrypsin deficiency associated nonalcoholic fatty liver disease, which may not be detected using the tradi- Cholestatic diseases (alkaline phosphatase and gamma-glutamyl transferase elevations tional, higher normal values. Acceptance is predominate) growing for using ALT levels less than 40 U/L in men and less than 31 U/L in women, and Common AST levels less than 37 U/L in men and less Primary biliary cirrhosis Primary sclerosing cholangitis than 31 U/L in women, as normal thresholds. Neoplasm Although ALT is present in several organs Biliary obstruction (gallstones, etc) and in muscle, the highest levels are in the Drug hepatotoxicity liver, which makes this enzyme a more specific indicator of liver injury. Both AST and ALT Nonalcoholic Less common Autoimmune cholangiopathy are released into the blood in greater amounts fatty liver Sarcoidosis when hepatocytes are damaged. disease Alkaline phosphatase suggests cholestasis may be the This article reviews the most likely causes Alkaline phosphatase comes mostly from the most common of elevated aminotransferase, alkaline phos- liver and bone. In general, normal serum al- phatase, and gamma-glutamyl transferase kaline phosphatase levels in adults range be- cause of mildly (GGT) levels. It also provides an algorithm tween 20 and 120 U/L. When bone disease elevated liver for evaluating mildly abnormal liver enzyme is excluded, an elevation suggests biliary ob- values in apparently healthy people. Patients struction, injury to the bile duct epithelium, enzymes with signs of hepatic decompensation need a or cholestasis. Additionally, there are rare cas- in US patients more concise and urgent evaluation. es of benign familial elevation of serum alka- line phosphatase, mainly of intestinal origin. ■■ PATTERNS OF LIVER ENZYME ELEVATION GGT is not specific “Liver function test” is commonly used to GGT is present in hepatocytes and biliary epi- describe liver enzyme measurements, but the thelial cells. The normal range is 0 U/L to 50 term should be reserved for tests of the func- U/L in men, and 0 U/L to 35 U/L in women. tional hepatic reserve—traditionally, the al- GGT elevation is the most sensitive marker bumin level and the prothrombin time.3 of hepatobiliary disease. However, its routine On the other hand, elevated serum liver clinical use is not recommended, as it cannot enzymes (aminotransferases, alkaline phos- by itself indicate a specific cause of liver dis- phatase, and GGT) can reflect abnormalities ease, although measuring the GGT level can in liver cells or in the bile duct. For example, help determine a hepatic origin for an isolated predominant elevation of aminotransfer- elevation of alkaline phosphatase.

196 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 Aragon and Younossi

■■ risk factors guide the workup northern European descent. In northern Eu- of ELEVATED eNZYMES rope, about 1 person in 10 is heterozygous and 1 in 200 to 400 is homozygous for the mutated Before beginning an extensive evaluation gene. of an elevated liver enzyme, a brief review Risk factors. Northern European ancestry of liver diseases and how they are associated is the primary risk factor. In men, the onset with specific liver enzyme elevations is useful of disease is usually in the third and fourth (Table 1). This information and clinical data decades of life, while menses protects women obtained from the history and physical exami- until menopause. From 83% to 85% of people nation provide important clues to guide fur- with clinically defined hemochromatosis are ther investigation. homozygous for the C282Y mutation in the HFE gene. Chronic viral hepatitis Comments. Hereditary hemochromato- Prevalence. Hepatitis C virus infection af- sis should be considered early in the evalu- fects an estimated 1.8% of the general popu- ation of men of northern European descent. lation, but the rate is much higher in people Patients usually have no symptoms until iron with known risk factors (see below), and those overload causes significant end-organ damage. with ALT levels greater than 40 U/L. Phlebotomy can be an effective treatment for Hepatitis B virus infection is somewhat this potentially fatal disease.6 less common: between 0.2% and 0.9% of the general US population have positive results Alcoholic liver disease on tests for hepatitis B surface antigen. How- Risk factors. The degree of alcohol-relat- ever, the prevalence of this antigen in the ed liver disease depends on a variety of factors, United States can be as high as 20% in pa- including the volume and duration of alcohol tients who have emigrated from endemic areas ingestion, the type of liver disease, genetics, of the world. The risk factors described below and the coexistence of viral hepatitis and obe- dramatically increase the prevalence of both sity. viruses. Alcohol-related liver disease can range Studies Risk factors. Risk factors include blood- from simple fatty liver to alcoholic hepatitis suggest that product transfusions (especially before 1992), with or without cirrhosis. Cirrhosis develops intravenous drug use, intranasal cocaine use, in only 20% to 30% of patients who consume fatty liver, due hemodialysis, organ transplantation, and birth a substantial amount of alcohol, defined as to alcohol or in an endemic region. Although both virus- more than a decade of 60 g/day to 80 g/day nonalcoholic es can be transmitted sexually, hepatitis B is of alcohol in men and as little as 20 g/day in more readily transmitted by this route than women. (A standard drink, ie, a 12-ounce steatohepatitis, hepatitis C. Worldwide, transmission of hepa- beer, a 5-ounce glass of wine, or 1.5 ounces is the major titis B virus usually occurs shortly after birth or of distilled spirits, contains 12 g of alcohol.) at a young age. Factors that potentiate alcohol’s harmful ef- cause of mildly Comments. Most patients with chronic fects include female sex, chronic viral hepa- elevated viral hepatitis have no symptoms or only titis (especially hepatitis C), obesity, heredi- mild symptoms and minimally elevated ALT tary hemochromatosis, and use of drugs such aminotransfer- and AST levels, ie, two to five times higher as methotrexate (Trexall) and acetaminophen ases than the upper limit of normal. Given the (Tylenol). relatively high prevalence of hepatitis C, se- Comments. Although cirrhosis affects few- rologic testing for it should be done early in er than one-third of long-term heavy drinkers, the evaluation of chronically elevated liver early detection and treatment can potentially enzyme levels.4,5 reduce morbidity and prevent early death. Al- coholic liver disease should be suspected in Hereditary hemochromatosis patients with elevated ALT and AST levels Prevalence. The prevalence of the major (if the AST level is two to three times higher HFE-gene mutations that cause hereditary he- than normal7) and with a history of excessive mochromatosis is 0.25% to 0.5% in people of alcohol use.

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 197 mildly Elevated liver enzymes

Nonalcoholic fatty liver disease is perhaps the Table 2 most common cause of mildly elevated liver Hepatotoxicity of selected drugs enzymes in the United States. Risk factors. The major risk factors for Hepatocellular abnormalities nonalcoholic fatty liver disease are the com- Acetaminophen (Tylenol)—acute hepatitis ponents of the metabolic syndrome—ie, ab- Allopurinol (Zyloprim)—granuloma dominal obesity, diabetes (insulin resistance), Azathioprine (Imuran)—veno-occlusive disease, hyperlipidemia, and hypertension—and the nodular regenerative hyperplasia use of certain medications (Table 2). Diclofenac (Voltaren) and other nonsteroidal anti-inflammatory drugs Comments. Nonalcoholic steatohepatitis Hydralazine (Apresoline)—granuloma and steatonecrosis describe a potentially pro- Isoniazid gressive form of nonalcoholic fatty liver dis- Methotrexate—fibrosis ease. Although these disorders are histologi- Methyldopa (Aldomet) cally indistinguishable from alcohol-induced Nitrofurantoin (Furadantin, Macrobid)— liver disease, their mechanism is related to in- autoimmune-like disease sulin resistance, abnormalities of lipid metab- Quinidine—granuloma olism, increased hepatic lipid peroxidation, Statins activated fibrocytes, and abnormal patterns Cholestatic abnormalities of adipokine and cytokine production related Amoxicillin-clavulanate (Augmentin) and other to visceral obesity. Results from a few natural penicillin derivatives history studies suggest that simple steatosis has Anabolic steroids a benign course, whereas nonalcoholic steato- Captopril (Capoten) hepatitis can progress to cirrhosis in 10% to Chlorpromazine (Thorazine) 20% of patients.8 Erythromycin estolate Treatment of diabetes, obesity, hyperten- Estrogens sion, and hyperlipidemia has potential benefit Oral contraceptives and should be undertaken regardless of liver Hepatobiliary Phenytoin (Dilantin)—mononucleosis-like test abnormalities in any patient with under- syndrome diseases and Carbamazepine (Tegretol) lying nonalcoholic fatty liver disease. Sulfa drugs several bone Autoimmune hepatitis Drug-induced fatty liver Prevalence. The prevalence of autoim- conditions (with or without hepatocellular abnormalities) can cause mune hepatitis varies, depending on geograph- Amiodarone (Cordarone)—phospholipidosis ic location and on the extent of viral hepatitis a moderately Corticosteroids in the community. In Hong Kong, only 1% of Tetracycline all people with chronic hepatitis have autoim- to markedly Valproic acid (Depakote) mune hepatitis. By contrast, in Germany and elevated Austria, 34% and 62% of patients with chron- alkaline ic hepatitis may have autoimmune hepatitis.9 Nonalcoholic fatty liver disease In North America, the prevalence of autoim- phosphatase Prevalence. Nonalcoholic fatty liver dis- mune hepatitis in patients with chronic liver level ease is a spectrum that ranges from simple ste- disease is estimated to be 11% to 23%, and the atosis to nonalcoholic steatohepatitis to cirrho- incidence is about 0.68 per 100,000 individu- sis. Its prevalence in the general US population als per year. In addition to underlying genetic is about 25%, but is much higher in groups at differences, detection bias can explain the risk, such as patients with type 2 diabetes (50% variability in prevalence rates. to 60%), and morbidly obese patients undergo- Risk factors. Autoimmune hepatitis oc- ing bariatric surgery (90% to 95%). curs predominantly in women and can be as- On the other hand, the prevalence of sociated with other autoimmune disorders. the potentially progressive form of nonal- Comments. The diagnosis of autoim- coholic fatty liver disease, ie, nonalcoholic mune hepatitis is suggested by exclusion of steatohepatitis, is estimated to be 3% to 5%. viral causes of chronic hepatitis, by pathologic

198 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 Aragon and Younossi

findings, and by the presence of autoimmune ment for the PiZZ genotype (the most severe markers such as antinuclear antibody, smooth form, because homozygous for the abnormal Z muscle antibody, and liver-kidney microsomal allele). antibody. Hypergammaglobulinemia is pres- Comments. Although alpha-1-antitrypsin ent in most patients, and serum protein elec- deficiency is a common cause of liver disease trophoresis may be helpful as part of the initial in the very young, it is important to remember evaluation of autoimmune hepatitis. that a small number of these patients develop Liver biopsy is usually needed to confirm end-stage liver disease in adulthood. Liver the diagnosis and to stage the extent of fibro- transplantation is the only effective treatment sis. The International Autoimmune Hepatitis for end-stage liver disease associated with al- Group Scoring System is based on clinical, pha-1-antitrypsin deficiency. laboratory, and pathologic data and can be very helpful in establishing the diagnosis. Primary biliary cirrhosis Treatment of autoimmune hepatitis with Prevalence. In one study of urban-dwelling immunosuppression is effective. Most patients women in northeast England, the prevalence may need long-term maintenance treatment. of primary biliary cirrhosis was estimated at 0.10%.13 Wilson disease Risk factors. Like autoimmune hepatitis, Prevalence. The estimated prevalence of primary biliary cirrhosis mainly affects women Wilson disease is 1 in 40,000 to 1 in 100,000. and can be associated with other autoimmune It has been reported in most populations disorders. worldwide. Comments. A cholestatic pattern of injury Risk factors. Anyone under age 40 with is predominant in primary biliary cirrhosis. abnormal liver enzyme levels (including mild Treatment of primary biliary cirrhosis with elevations) should be evaluated for Wilson the cytoprotective agent ursodeoxycholic acid disease, even in the absence of neurologic improves liver enzyme levels, may lead to his- or ocular findings. However, such routine tologic improvement and increased survival, screening is rarely helpful in patients over and may also delay the need for liver trans- Extensive age 50. Genetic testing is of limited value plantation.9,13 workup because of the large number of potential mu- tations of the ATP7B gene, the gene respon- Drug- and toxin-related liver diseases of an isolated sible for Wilson disease. However, if a a per- Nonsteroidal anti-inflammatory drugs and GGT elevation son is known to have Wilson disease, genetic penicillin-derived antibiotics are the drugs in an screening of family members is useful. that most commonly cause abnormal serum Comments. Effective therapy is available liver enzyme levels. The mechanisms of drug- apparently (ie, d-penicillamine, trientine, zinc). For Wil- induced liver disease include induction of he- healthy son disease, alpha-1-antitrypsin deficiency, patic enzymes (antiepileptic drugs), allergic and genetic hemochromatosis, establishing reactions, autoimmunity (nitrofurantoin [Fu- patient is not the diagnosis is not only important to the indi- radantin, Macrobid]), idiosyncratic reactions, warranted vidual patient; it also may prompt the screen- and veno-occlusive injury.14 Drugs that are po- ing of asymptomatic members of the proband’s tentially hepatotoxic are listed in Table 2 and family.10–12 are classified as causing hepatocellular dam- age, cholestatic damage, or steatosis. Alpha-1-antitrypsin deficiency Prevalence. Alpha-1-antitrypsin deficien- ■■ MILD ENZYME ELEVATIONS cy is present in 1 of every 1,600 to 1,800 live AS INDICATORS OF SPECIFIC DISEASES births.11 Risk factors. Patients with emphysema or Mildly elevated liver enzymes are common with a young sibling with liver failure should and potentially important, yet very few well- undergo an investigation for alpha-1-antitryp- designed prospective studies have addressed sin deficiency, consisting of a measurement of the issue of what should be done once they the alpha-1-antitrypsin level and an assess- are identified. Most current data are from

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 199 mildly Elevated liver enzymes

small retrospective studies that lack accurate mildly elevated aminotransferase levels: information on the important causes of liver Hultcrantz et al1 performed a full evalua- diseases such as hepatitis C and nonalcoholic tion, including liver biopsy, in 149 consecutive steatohepatitis. patients with chronic, asymptomatic, mild el- Despite these shortcomings, the literature evations of AST or ALT. Of these patients, delineates the three patterns of mild liver en- 63% had “fatty liver,” 20% had “chronic hepa- zyme elevations discussed earlier: hepatocellular titis,” and 17% had miscellaneous diagnoses. injury pattern (elevated ALT or AST), chole- Whether patients in the “chronic hepatitis” static pattern (elevated alkaline phosphatase or group had hepatitis C was not determined be- GGT, or both), and mixed pattern (elevation of cause serologic testing was not available at the ALT, AST, and alkaline phosphatase). The fol- time. lowing paragraphs focus on the causes of eleva- Friedman et al16 studied 100 healthy blood tion of specific liver enzymes. donors with elevated ALT levels and found that in 33% of patients the elevation occurred ■■ Aminotransferase elevation once, in 36% it was intermittent, and in 28% it was persistent. In this series, 45% of patients Causes had no diagnosis, 22% were obese (presumed Aminotransferases are commonly used mark- to have nonalcoholic steatohepatitis), 5% ers of hepatocyte injury. AST is present in had alcoholic liver disease, 3% had “resolving blood cells and many tissues, including liver, hepatitis,” 1% had hemochromatosis, and 1% muscle, brain, pancreas, and lung. ALT is a had “cytomegalovirus hepatitis.”16 Although cytosolic enzyme found primarily in hepato- the patients underwent a complete history, cytes, making it a more specific indicator of physical examination, and serologic testing, liver disease. liver biopsy was not done to confirm the clini- Acute viral hepatitis, toxins, and liver cal diagnosis. ischemia can markedly raise serum amino- Hay et al17 described 47 patients with transferase levels (often into the thousands of chronically elevated aminotransferases (three An AST level units per liter). On the other hand, these en- to eight times higher than normal levels) who more than zymes are only mildly elevated (< 300 U/L) in underwent full evaluation and liver biopsy and nonalcoholic steatohepatitis, chronic hepati- who had no clinical symptoms of alcoholic, vi- twice the ALT tis, cholestatic liver conditions, drug-induced ral, or drug-induced liver disease. A diagnosis suggests the hepatotoxicity, and liver tumors. of steatohepatitis was given in 10 patients, an- damage is due Because AST is a mitochondrial enzyme other 34 were diagnosed with “chronic hepati- and is affected by alcohol ingestion, an AST tis,” and 3 had miscellaneous diagnoses. Of pa- to alcohol level more than twice that of the ALT suggests tients with chronic hepatitis, 16 had evidence hepatic damage due to alcohol. of cirrhosis on biopsy, and 18 tested positive for Of note, aminotransferase elevation can at least one autoimmune marker (antinuclear also be due to nonhepatic causes. For exam- antibody or smooth muscle antibody). ple, muscle necrosis can result in mild eleva- Daniel et al18 performed biopsy in 81 of tion of these enzymes, especially AST, and an 1,124 asymptomatic and symptomatic pa- elevated creatine kinase can help confirm that tients with chronically elevated aminotrans- the source is muscle tissue. ferases in whom a cause was not identified via Undiagnosed celiac disease has been as- noninvasive studies. Liver biopsy showed that sociated with abnormal liver enzyme levels 67 (83%) of the 81 patients had steatosis or when all other causes have been ruled out, steatohepatitis, while 8 (10%) had normal but the mechanism is not yet understood. In histologic findings. Of note, 6 patients had this situation, ALT and AST levels typically underlying fibrosis or cirrhosis and some de- return to normal with a gluten-free diet.15 gree of fatty infiltration. Together, these studies suggest that fatty Studies of mild aminotransferase elevations liver, resulting either from alcohol use or from Only a few studies have documented the re- nonalcoholic fatty liver disease, is the major sults of a thorough evaluation of patients with cause of mildly elevated aminotransferases.

200 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 Aragon and Younossi

Two major drawbacks of the earlier studies hemochromatosis). Iron studies for middle- include the lack of data on the hepatitis C se- aged men, autoimmune markers for women, rologic status of patients with the diagnosis of and screening for Wilson disease in young “chronic hepatitis” and the lack of a uniform patients are helpful when the clinical infor- approach to the pathologic diagnosis of non- mation points to one of these entities as a alcoholic steatohepatitis. With serologic test- potential diagnosis. ing for hepatitis C virus infection now widely If medication or alcohol is a suspected cause, available, it is possible that a substantial por- aminotransferase levels should be repeated af- tion of patients with “chronic hepatitis“ can ter 6 to 8 weeks of abstinence. If nonalcoholic further be classified as having chronic hepati- steatohepatitis is suspected, testing should be tis C infection. repeated after treating the potential risk factor (eg, obesity, diabetes, hyperlipidemia), but the Workup of aminotransferase elevations levels may remain elevated for a period of time. Figure 1 shows an algorithm for evaluating pa- An imaging study (ultrasonography, computed tients with elevated aminotransferase levels tomography, or magnetic resonance imaging) on an initial examination. The first step is may be helpful; eg, abdominal ultrasonography to confirm the abnormality by repeating the may show increased hepatic echogenicity, sug- blood test. If an enzyme elevation is confirmed, gesting increased fatty infiltration, in addition further investigation is warranted. A directed to excluding most hepatic tumors. history and physical examination can provide If the clinical data obtained from the his- crucial clues in the preliminary workup. The tory or physical examination raise clinical sus- history may disclose risk factors for: picion for a particular disease, a disease-spe- • Viral hepatitis (intravenous drug use, in- cific marker (figure 1 and figure 2) can further tranasal cocaine use, native of an endemic support the potential diagnosis. Note that liv- area of the world, unsafe sexual activity, er biopsy can help establish the diagnosis for blood product transfusions) many liver disorders and is the best method • Nonalcoholic fatty liver disease (compo- currently available to establish cirrhosis, with nents of the metabolic syndrome, includ- important prognostic implications.19 If a drug ing visceral obesity) If the history and physical do not sug- or alcohol is • Alcoholic liver disease (smaller amounts gest a specific condition, serologic testing for are needed to cause liver disease in women) hepatitis C should be done. If negative, other the suspected • Medication exposure (prescription, over- testing can be helpful (iron studies in men, cause, retest the-counter, and herbal medications) autoimmune markers in women, ceruloplas- AST and ALT • Genetic liver disorders (family history of min and slit-lamp examination in young pa- liver disease) tients). If the preliminary workup remains after 6 to 8 • Possible coexisting disease (diabetes and negative and the aminotransferase levels re- weeks of obesity in nonalcoholic steatohepatitis, main elevated for 6 months (ie, chronic el- neurologic disorders in Wilson disease, evation), liver biopsy may establish the diag- abstinence emphysema in alpha-1-antitrypsin defi- nosis. Features in the biopsy specimens may ciency, thyroid disease in autoimmune further confirm the diagnosis: eg, globules hepatitis and primary biliary cirrhosis, and positive on periodic acid-Schiff testing in diabetes and impotence in genetic hemo- alpha-1-antitrypsin deficiency; hepatic iron chromatosis). index for hemochromatosis; hepatic copper Although the physical signs of chronic content for Wilson disease. liver disease (eg, spider angiomata, palmar erythema, caput medusae, and gynecomas- ■■ Alkaline phosphatase elevation tia) are nonspecific and are usually observed in advanced liver disease, some physical Causes findings suggest potential causes (eg, Kayser- Alkaline phosphatase is active in many or- Fleischer rings on slit-lamp examination for gans, mainly the liver and bones, but is also Wilson disease, hypertrophy of the second found in the small bowel, kidneys, and pla- and third metacarpophalangeal joints for centa. Diseases of the hepatobiliary system

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 201 mildly Elevated liver enzymes

Elevated alanine or aspartate aminotransferase

History and physical examination Retest to confirm elevation

Normal Abnormal

Test for hepatitis C Disease suspected antibody if at risk (risk factors present) + Nonalcoholic fatty Autoimmune Wilson disease Hemochromatosis Viral hepatitis Alcohol- or Hepatitis C virus liver disease hepatitis drug-related RNA + Treat underlying risk Test positive for Tests: Tests: Test for hepatitis Recheck after factors antinuclear anti- Ceruloplasmin Ferritin C virus, hepatitis 6-8 weeks of body, anti-smooth- B surface antigen abstinence muscle antibody 24-hour urine Percent iron saturation Chronic Slit-lamp test for + hepatitis C Kayser-Fleischer HFE mutation rings Hepatitis C virus RNA or hepatitis B virus DNA

Abnormal Normal Consider liver biopsy Follow-up figure 1. Algorithm for the evaluation of elevated aminotransferase levels. Black arrows indicate the diagnostic pathway, and red arrows indicate the steps in staging the liver disease.

can cause moderate to marked elevations and primary sclerosing cholangitis) of alkaline phosphatase. Conditions with • Infiltrative process (eg, neoplasm, tubercu- bone involvement, such as Paget disease of losis, sarcoidosis) the bone, sarcoma, metastatic disease, hy- • Cholestatic disorders (eg, drug hepato- perparathyroidism, and rickets, can raise al- toxicity) kaline phosphatase levels. Elevated GGT in • Biliary obstruction (eg, in neoplasia or conjunction with elevated alkaline phospha- cholelithiasis). tase usually points to hepatobiliary injury. Only a few studies have investigated the Clinically, isoenzyme fractionation of alka- significance of a mild, isolated elevation of al- line phosphatase may help distinguish the kaline phosphatase. Lieberman and Phillips20 source of the elevation, but this is often not evaluated 87 patients and found that the ab- needed if the GGT is also elevated. normality resolved completely in less than 3 Hepatobiliary causes of alkaline phospha- months in 28 patients, and resolved in 3 to 12 tase elevation can be divided into four catego- months in another 17 patients. Of the other ries: 42 patients, 24 did not undergo further evalu- • Chronic inflammation involving the bile ation because they had significant coexisting ducts (eg, as in primary biliary cirrhosis disease. Of the remaining 18 patients, 5 had

202 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 Aragon and Younossi

Elevated alkaline phosphatase

History and physical examination Consider magnetic resonance cholangiopancreatography Confirm and identify hepatobiliary source, check gamma-glutamyl transferase to exclude alcohol- or drug-related hepatitis and biliary discomfort Consider expert consultation Consider Biliary and endoscopic retrograde Obstruction ultrasonography dilation cholangiopancreatography

Treat underlying Abnormal cause of obstruction

Alkaline phosphatase still Disease suspected elevated at 6 months if risk factors present

Primary sclerosing Viral hepatitis Primary biliary Drug-related cholangitis cirrhosis Test for hepatitis C virus, Recheck after Cholangiopancreatography hepatitis B surface antigen Test for antimito- 6-8 months of (magnetic resonance or endo- (for mixed pattern of injury) chondrial antibody abstinence scopic retrograde)

Abnormal Normal Consider liver biopsy Follow-up figure 2. Algorithm for the evaluation of elevated alkaline phosphatase levels. Black arrows indicate the diagnostic pathway, and red arrows indicate the steps in staging the liver disease. phenytoin-related hepatotoxicity, 3 had con- tion from the history can point to a potential gestive heart failure, 3 had metabolic bone underlying pathologic process causing the disease, 2 had hepatobiliary disease, and 1 had rise in alkaline phosphatase. For example, a metastatic bone disease; in 4 patients, no ex- history of ulcerative colitis suggests primary planation was determined. Follow-up was 1.5 sclerosing cholangitis, and a history of pre- to 3 years. vious cancer or sarcoidosis suggests liver in- volvement. Workup of alkaline phosphatase elevation As part of the initial evaluation, an imag- An isolated elevated alkaline phosphatase ing study (eg, ultrasonography) can exclude level should always be confirmed and a he- biliary obstruction or suggest an infiltrative patic origin suspected if the GGT level is also process. elevated (Figure 2). If a drug is the suspected cause, the al- A history of recent drug exposure usually kaline phosphatase level should be repeated points to drug hepatotoxicity as the source after the patient has abstained from the drug of this abnormality. Similarly, other informa- for 6 to 8 weeks. If the initial examination

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010 203 mildly Elevated liver enzymes

suggests a specific disease, disease-specific without known liver disease. The GGT level markers (eg, antimitochondrial antibody for can be used to monitor abstinence from alcohol primary biliary cirrhosis, or viral serology) in patients with alcoholic liver disease.21 can confirm the suspected diagnosis. If the disease-specific markers are negative and the Workup of GGT elevation alkaline phosphatase level does not return to Because GGT lacks specificity as a marker and normal, further studies should be considered, is highly inducible, an extensive evaluation including liver biopsy and endoscopic retro- of an isolated GGT elevation in an otherwise grade cholangiopancreatography or magnetic asymp­tomatic patient is not warranted. resonance cholangiopancreatography. Given the invasive nature of biopsy and pancreatog- ■■ When to consult raphy, an expert consultation should be done before ordering these tests. Many of the evaluations discussed in this paper and elsewhere22–27 can be carried out by primary ■■ GGT elevation care providers following a systematic approach. Input from a gastroenterologist or hepatolo- GGT is a membrane enzyme that is a marker of gist can be valuable if the initial workup fails hepatobiliary disease. Elevations usually parallel to establish the diagnosis, as well as in assur- the elevation of alkaline phosphatase, confirm- ing that the most effective therapy for a specific ing a hepatic source for the latter. GGT is the disease is initiated. Reassurance, patient educa- most sensitive marker of biliary tract disease tion, and a systematic approach for evaluating but is not very specific. Alcohol and drugs (eg, these abnormalities can identify most treatable phenytoin, pheno­barbital) induce GGT. In one causes of liver disease in the most cost-effective study,21 GGT was elevated in 52% of patients and efficient manner. ■ ■■ REFERENCES 15. abdo A, Meddings J, Swain M. Liver abnormalities in celiac disease. Clin Gastroenterol Hepatol 2004; 2:107–112. 1. Hultcrantz R, Glaumann H, Lindberg G, Nilsson LH. Liver investigation 16. Friedman LS, Dienstag JL, Watkins E, et al. Evaluation of blood donors in 149 asymptomatic patients with moderately elevated activities of with elevated serum alanine aminotransferase. Ann Intern Med 1987; serum aminotransferases. Scand J Gastroenterol 1986; 21:109–113. 107:137–144. 2. ioanou GN, Boyko EJ, Lee SP. The prevalence and predictors of 17. Hay JE, Czaja AJ, Rakela J, Ludwig J. The nature of unexplained elevated serum aminotransferase activity in the United States in 1999- chronic aminotransferase elevations of a mild to moderate degree in 2002. Am J Gastroenterol 2006; 101:76–82. asymptomatic patients. Hepatology 1989; 9:193–197. 3. Flora KD, Keeffe EB. Evaluation of mildly abnormal liver tests in 18. Daniel S, Ben-Menachem T, Vasudevan G, Ma CK, Blumenkehl M. asymptomatic patients. J Insur Med 1990; 22:264–267. Prospective evaluation of unexplained chronic liver transaminase 4. alter HJ. To C or not to C: these are the questions. Blood 1995; abnormalities in asymptomatic and symptomatic patients. Am J Gas- 85:1681–1695. troenterol 1999; 94:3010–3014. 5. everhart JE, editor. Digestive diseases in the United States: epidemiol- 19. Van Ness MM, Diehl AM. Is liver biopsy useful in the evaluation of pa- ogy and impact. Washington, DC: US Government Printing Office, tients with chronically elevated liver enzymes? Ann Intern Med 1989; 1994; NIH Publication No. 94-1447. 111:473–478. 6. bacon BR, Tavill AS. Hemochromatosis and the iron overload syn- 20. lieberman D, Phillips D. “Isolated” elevation of alkaline phosphatase: sig- dromes. In: Zakim D, Boyer TD, editors. Hepatology. A Textbook of nificance in hospitalized patients. J Clin Gastroenterol 1990; 12:415–419. Liver Disease, 3rd ed. Philadelphia: WB Saunders, 1996:1439–1472. 21. Margarian GJ, Lucas LM, Kumar KL. Clinical significance in alcoholic 7. lieber CS. Alcoholic liver disease. Curr Opin Gastroenterol 1994; patients of commonly encountered laboratory test results. West J Med 10:319–330. 1992; 156:287–294. 8. Seth SG, Gordon FD, Chopra S. Nonalcoholic steatohepatitis. Ann 22. Fregia A, Jensen DM. Evaluation of abnromal liver tests. Compr Ther Intern Med 1997; 126:137–145. 1994; 20:50–54. 9. Czaja A. Autoimmune liver disease. In: Zakim D, Boyer TD, editors. 23. goddard CJ, Warens TW. Raised liver enzymes in asymptomatic pa- Hepatolology: A Textbook of Liver Disease, 3rd ed. Philadelphia: WB tients: investigation and outcome. Dig Dis 1992; 10:218–226. Saunders, 1996:1259–1292. 24. gitlin N. The differential of elevated liver enzymes. Contemporary 10. bull PC, Cox DW. Wilson disease and Menkes disease: new handles on Internal Medicine 1993:44–56. heavy-metal transport. Trends Genet 1994; 10:246–252. 25. Keefe EB. Diagnostic approach to mild elevation of liver enzyme 11. Perlmutter DH. Clinical manifestations of alpha-1-antitrypsin defi- levels. Gastrointestinal Diseases Today 1994; 3:1-9. ciency. Gastroenterol Clin North Am 1995; 24:27–43. 26. Pratt DS, Kaplan MM. Evaluation of abnormal liver-enzyme results in 12. Olivarez L, Caggana M, Pass KA, Ferguson P, Brewer GJ. Estimate of asymptomatic patients. N Engl J Med 2000; 342:1266–1271. the frequency of Wilson’s disease in the US Caucasian population: a 27. american Gastroenterological Association. American Gastroentero- mutation analysis approach. Ann Hum Genet 2001; 65:459–463. logical Association medical position statement: evaluation of liver 13. Metcalf JV, Howel D, James OF, Bhopal R. Primary biliary cirrhosis: chemistry tests. Gastroenterology 2002; 123:1364–1366. epidemiology helping the clinician. BMJ 1996; 312:1181–1182. 14. Farrell GC. Liver disease caused by drugs, anesthetics, and toxins. In: ADDRESS: Zobair M. Younossi, MD, MPH, Center for Liver Diseases, Inova Feldman M, Friedman LS, Sleisenger MH, editors. Gastrointestinal and Fairfax Hospital, 3300 Gallows Road, Falls Church, VA 22042; e-mail zobair. Liver Disease, 7th ed. Philadelphia: WB Saunders, 2002; 1403–1447. [email protected].

204 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 77 • NUMBER 3 MARCH 2010