DIAGNOSTIC APPROACH TO JAUNDICE

A general algorithm for evaluating the patient with jaundice is depicted in Figure 20-2. The sequential approach involves the following: (1) a carefully taken patient history, thorough physical examination, and screening laboratory studies; (2) formulation of a working differential diagnosis; (3) selection of specialized tests to narrow the diagnostic possibilities; and (4) development of a strategy for treatment or further testing if unexpected diagnostic possibilities arise.
Figure 20-2. Algorithm for the evaluation and management of jaundice and hyperbilirubinemia. CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasound; MRCP, magnetic resonance cholangiopancreatography; THC, transhepatic cholangiography.

HISTORY AND PHYSICAL EXAMINATION

The patient's history and physical examination provide important clues regarding the cause of jaundice (Table 20-3). A history of biliary surgery, fever, especially when accompanied by rigors, and abdominal pain, particularly in the right upper quadrant, is suggestive of biliary obstruction with cholangitis. Obstructive jaundice from gallstone disease or malignant neoplasms is more common in older adults than in younger persons. Symptoms compatible with a viral prodrome, such as anorexia, malaise, and myalgias, make viral hepatitis a strong diagnostic possibility, as does a history of a known infectious exposure, injection drug use, or prior transfusions of blood products. A carefully taken history may suggest that environmental hepatotoxins, ethanol, or medications underlie the patient's cholestatic liver disease. Furthermore, a family history of jaundice or liver disease suggests the possibility of hereditary hyperbilirubinemia or genetic liver disease. All clues must be interpreted with caution, because fever and abdominal pain accompany diseases other than biliary obstruction, and viral hepatitis may occur coincidentally in patients with a history of prior biliary surgery. Conversely, anorexia and malaise are not specific for viral hepatitis, and gallstones frequently develop in patients with chronic liver disease. Nonetheless, when clues are evaluated in the context of the physical findings and routine laboratory tests, jaundice can be characterized correctly as [42] obstructive or nonobstructive in at least 75% of cases. Table 20-3 -- Clues to the Differential Diagnosis of Jaundice: Biliary Obstruction versus Liver Disease PARAMETER BILIARY OBSTRUCTION LIVER DISEASE Anorexia, malaise, myalgias (viral prodrome) Known viral exposure History of blood product receipt or of injection drug use Exposure to known hepatotoxin Family history of liver disease Fever Physical examination Abdominal tenderness Palpable abdominal mass Abdominal surgical scar Laboratory studies Predominant elevation of serum Spider angiomata Stigmata of portal hypertension (e.g., prominent abdominal veins, splenomegaly, ascites) Asterixis Predominant elevation of serum

Abdominal pain Fever, rigors History Prior biliary surgery Older age

PARAMETER BILIARY OBSTRUCTION alkaline phosphatase relative to aminotransferases* Prothrombin time (INR) normal or normalizes with vitamin K administration Leukocytosis Elevated serum amylase or lipase level INR, international normalized ratio.
*

LIVER DISEASE aminotransferase levels relative to alkaline phosphatase Prolonged prothrombin time that does not normalize with vitamin K administration Thrombocytopenia Serologies indicative of specific liver disease

Except early after acute obstruction when the opposite pattern may be seen transiently.

The clues offered by the physical examination also are important in the patient with jaundice. High fever or abdominal tenderness (particularly in the right upper quadrant) suggests cholangitis, and a palpable abdominal mass suggests a neoplastic cause of obstructive jaundice. The rare finding of silver stools, resulting from the combination of blood and lack of bile, suggests an ampullary neoplasm. An abdominal scar in the midline or right upper quadrant may be the only clinical clue to prior biliary surgery. The presence of cirrhosis may be suggested by signs of portal hypertension, such as ascites, splenomegaly, and prominent abdominal veins, or other physical findings of liver disease, such as spider angiomata, gynecomastia, and asterixis. Certain physical findings may suggest specific liver diseases, as for hyperpigmentation in hemochromatosis, xanthomas in primary biliary cirrhosis, and Kayser-Fleischer rings in Wilson disease.

INITIAL LABORATORY STUDIES

Essential laboratory studies in the patient with jaundice include serum total bilirubin, alkaline phosphatase, ALT, and AST levels, a complete blood count, and the prothrombin time (see Chapter 73). Serum alkaline phosphatase activity reflects a number of related enzymes of overlapping substrate specificity. Alkaline phosphatase is associated predominantly with the apical domain of the plasma membrane of hepatocytes and cholangiocytes. Under physiologic conditions, this protein is cleaved enzymatically from a glycolipid anchor and released into bile, and small amounts are released from the sinusoidal (basolateral membrane) into plasma as well. Biliary obstruction and intrahepatic cholestasis increase the synthesis and basolateral release of alkaline phosphatase, and serum alkaline phosphatase activity increases. An increase in serum alkaline phosphatase activity, however, also may reflect release of alkaline phosphatase isoenzymes from extrahepatic tissues. Therefore, other more specific markers, such as the serum activities of the canalicular enzymes gamma glutamyl transpeptidase, or 5 nucleotidase (or alternatively, alkaline phosphatase isoenzymes), are measured to confirm the hepatobiliary origin of an elevated serum alkaline phosphatase level when other liver biochemical test results (e.g., total bilirubin, ALT, AST) are normal. In a jaundiced patient, a predominant increase in (hepatic) alkaline phosphatase activity relative to levels of the serum aminotransferases suggests the possibility of biliary tract obstruction. Intrahepatic cholestatic disorders can produce an identical biochemical picture. The aminotransferasesALT, a cytosolic enzyme found predominantly in hepatocytes, and AST, isozymes of which are found in both the cytosol and mitochondria of parenchymal cells of liver and several other tissuesare ordinarily detected in serum in low concentrations. Hepatocellular injury caused by ischemia, toxins, or immune-mediated responses to foreign antigens such as viral proteins greatly increases serum aminotransferase activity. Predominant elevation of serum aminotransferase activity in comparison with alkaline

phosphatase activity suggests that jaundice is the result of intrinsic hepatocellular disease. A serum activity of AST that is less than 10 times the upper limit of normal and that exceeds ALT activity by at least a factor of 2 is usually suggestive of alcoholic liver disease (see Chapter 84), but there are exceptions to these generalizations. For example, transient biliary obstruction from choledocholithiasis associated with cholangitis may cause a brief but [43] dramatic elevation (exceeding 10 to 20 times normal) of serum aminotransferase activity. A complete blood count provides complementary information concerning the cause of jaundice. Leukocytosis may be a clue to the presence of biliary tract obstruction or other inflammatory disorder that may be associated with cholestasis. The presence of anemia leaves open the possibility that a hemolytic disorder is responsible for bilirubin overload. Thrombocytopenia is a characteristic finding in cirrhosis and appears to result from reduced platelet production from decreased hepatocyte synthesis of thrombopoietin or from increased platelet consumption from splenic sequestration associated with portal hypertension. The prothrombin time is a measure of the plasma activities of coagulation factors I, II, V, VII, and X, each of which is synthesized by hepatocytes. Prolongation of the prothrombin time (and an associated increase in the international normalized ratio, INR) can result from impaired hepatic synthesis of these proteins and from deficiency of vitamin K, which is required as a cofactor for essential posttranslational modification of factors II, VII, IX, and X. Efficient absorption of vitamin K by the small intestine requires an intact enterohepatic circulation of bile salts (hence, an unobstructed biliary tree). Exogenous administration of vitamin K will generally normalize a prolonged prothrombin time in patients with obstructive jaundice and intrahepatic cholestasis but not in patients with liver disease caused by hepatocellular injury.

OVERALL APPROACH
Integration of the patient's history, physical examination, and laboratory study results will provide an estimate of the likelihood that obstructive jaundice is present. For example, an asymptomatic patient with hyperbilirubinemia who has an unremarkable physical examination, normal serum alkaline phosphatase and aminotransferase levels, normal platelet count, and normal prothrombin time is unlikely to have liver disease or biliary obstruction. In this patient, further testing for specific disorders, such as isolated defects in bilirubin metabolism or hemolysis, is warranted (see Fig. 20-2). Alternatively, if the history, physical examination, and laboratory study results suggest the possibility of biliary obstruction, an imaging study of the biliary tree is appropriate. Selection of the appropriate imaging study depends on the likelihood of bile duct obstruction and the diagnostic accuracy, cost, complication rate, and availability of each test (see later), especially if therapeutic intervention at the time of the study is anticipated.

IMAGING STUDIES Abdominal Ultrasonography

Abdominal ultrasonography is usually the initial imaging test in the evaluation of hepatobiliary disease because it determines the caliber of the extrahepatic biliary tree and reveals intra- or extrahepatic mass lesions. The sensitivity of abdominal ultrasonography for the detection of biliary obstruction in jaundiced patients ranges from 55% to 91%, and the specificity ranges [44-48] from 82% to 95%. Ultrasonography also can demonstrate cholelithiasis, although bile duct stones may not be well seen, and intrahepatic space-occupying lesions more than 1 cm in diameter. Ultrasonography has the advantages of being noninvasive, portable (invaluable in the evaluation of the critically ill patient), and relatively inexpensive (Table 20-4). The major disadvantages are that the procedure is operator-dependent and interpretation may be difficult in obese patients or patients with overlying bowel gas. An additional caveat is that in patients with cirrhosis and other conditions associated with poorly compliant hepatic parenchyma, such as primary sclerosing cholangitis, intrahepatic ducts may not dilate with biliary obstruction. Table 20-4 -- Imaging Studies for the Evaluation of Jaundice

TEST

SENSITIVITY SPECIFICITY MORBIDITY MORTALITY ADVANTAGES AND (%) (%) (%) (%) DISADVANTAGES Advantages noninvasive, portable

Abdominal US

55-91

82-95

Disadvantages bowel gas may obscure bile duct; difficult in obese persons, operatordependent Advantages noninvasive, higher resolution than ultrasound, not operator-dependent Disadvantages potential for contrastinduced nephrotoxicity, anaphylaxis Advantages noninvasive, imaging of bile ducts superior to ultrasound and CT Disadvantages requires breath holding, may miss small-caliber bile duct disease Advantages provides direct imaging of bile ducts; permits direct visualization of periampullary region and acquisition of tissue distal to bifurcation of hepatic ducts; permits simultaneous therapeutic intervention, especially useful for lesions distal to bifurcation of hepatic ducts Disadvantages requires sedation, cannot be performed

Abdominal CT

63-96

93-100

See disadvantages

MRCP

82-100

94-98

See disadvantages

ERCP

89-98

89-100

0.2

TEST

SENSITIVITY SPECIFICITY MORBIDITY MORTALITY ADVANTAGES AND (%) (%) (%) (%) DISADVANTAGES if altered anatomy precludes endoscopic access to ampulla (e.g., Roux-en-Y loop); has complications (e.g., pancreatitis) Advantagesprovides direct imaging of bile ducts, permits simultaneous therapeutic intervention, especially useful for lesions proximal to common hepatic duct Disadvantagesmore difficult with nondilated intrahepatic bile ducts; has complications Advantagesimaging of bile ducts superior to ultrasound and CT, permits needle aspiration of suspected neoplasms Disadvantages requires sedation

Percutaneous 98-100 THC

89-100

3.5

0.2

EUS

89-97

67-98

See disadvantages

CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasonography; MRCP, magnetic resonance cholangiopancreatography; THC, transhepatic cholangiography; US, ultrasonography.

Computed Tomography
Computed tomography (CT) of the abdomen with intravenous contrast is an alternative noninvasive means of evaluating the possibility of biliary tract obstruction. Abdominal CT permits accurate measurement of the caliber of the biliary tree, with sensitivity and specificity rates of 63% to 96% and 93% to 100%, respectively, for detecting biliary obstruction; these [44-47] rates are comparable with those for ultrasonography. Abdominal CT detects intrahepatic space-occupying lesions as small as 5 mm, is not operator-dependent, and provides technically superior images in obese persons and in those in whom the biliary tree is obscured by bowel gas. The caveats that apply to the accuracy of ultrasonography for the diagnosis of biliary obstruction also apply to abdominal CT. Abdominal CT also lacks portability, it is more expensive than ultrasonography, and the requirement for the use of intravenous contrast is a potential contraindication in the setting of kidney failure (see Table 20-4).

Magnetic Resonance Cholangiopancreatography

Magnetic resonance cholangiopancreatography (MRCP) is a technical refinement of standard magnetic resonance imaging that permits rapid clear-cut delineation of the biliary tree without the need for intravenous contrast. MRCP appears to be superior to conventional ultrasound or CT for the detection of biliary tract obstruction and now plays a major role as a diagnostic test in this setting (see Table 20-4). Moreover, standard magnetic resonance imaging can be performed during the same examination if there is a question of a hepatobiliary or pancreatic

mass or if a contrast allergy precludes CT. For detection of obstruction of the bile ducts, the [49-52] sensitivity of MRCP is 82% to 100% and the specificity is 94% to 98%. Its expense is [53] higher than that of ultrasound or CT and comparable with that of ERCP.

Endoscopic Retrograde Cholangiopancreatography

Endoscopic retrograde cholangiopancreatography (ERCP) permits direct visualization of the biliary tree. ERCP is more invasive than ultrasonography and CT (see Table 20-4). The procedure involves passage of an endoscope into the duodenum, introduction of a catheter into the ampulla of Vater, and injection of contrast medium into the bile duct; sedation and analgesia are necessary. ERCP is highly accurate in the diagnosis of biliary obstruction, with [47,54,55] sensitivities of 89% to 98% and specificities of 89% to 100%. In addition to providing radiographic images, ERCP permits biopsy and brushings for cytology of distal biliary and periampullary lesions. Moreover, if a focal cause of biliary obstruction is identified (e.g., choledocholithiasis, biliary stricture), maneuvers to relieve obstruction (e.g., sphincterotomy, stone extraction, stricture dilation, stent placement) can be performed during the same session (see Chapter 70). Acquisition of biopsy specimens and therapeutic interventions via ERCP are limited largely to lesions distal to the bifurcation of the right and left hepatic bile ducts. The technical success rate of ERCP is higher than 90%; the technique fails when the ampulla of Vater cannot be cannulated, as may be the case in patients with prior abdominal surgery and altered anatomy (e.g., gastric bypass, choledochojejunostomy). Rates of morbidity and mortality from untoward events, such as respiratory depression, aspiration, bleeding, perforation, cholangitis, and pancreatitis, are 3% and 0.2%, respectively, in patients [55] undergoing ERCP. These rates are higher when interventional procedures are carried [56] out.

Percutaneous Transhepatic Cholangiography

Percutaneous transhepatic cholangiography (THC) is a procedure that complements ERCP. Percutaneous THC requires the passage of a needle through the skin and subcutaneous tissues into the hepatic parenchyma and advancement into a peripheral bile duct. When bile is aspirated, a catheter is introduced through the needle, and radiopaque contrast medium is injected. Sensitivity and specificity rates of percutaneous THC for the diagnosis of biliary tract obstruction are 98% to 100% and 89% to 100%, respectively, and are comparable with those [57,58] for ERCP. Like ERCP, interventional procedures, such as balloon dilation and stent placement, can be performed at the time of percutaneous THC to relieve focal obstructions of the biliary tree (see Chapter 70). Percutaneous THC is potentially technically advantageous when the level of biliary obstruction is proximal to the common hepatic duct or altered anatomy precludes ERCP (see earlier). Percutaneous THC may be technically challenging in the absence of dilatation of the intrahepatic bile ducts; in this situation, multiple passes may be required, and visualization of the biliary tree may be unsuccessful in up to 10% of [59] attempts. Rates of morbidity and mortality as a result of bleeding, perforation, and cholangitis are 3% and 0.2%, respectively, in patients undergoing percutaneous THC. Percutaneous THC is more expensive than abdominal ultrasonography and CT (see Table 204).

Endoscopic Ultrasonography
Endoscopic ultrasonography (EUS) also can detect obstruction of the bile duct and major intrahepatic bile ducts, with a sensitivity and specificity comparable with those of [49,60,61] MRCP. EUS has the potential advantage of permitting biopsy of suspected malignant lesions, and under appropriate circumstances, the operator can proceed directly to ERCP for definitive biliary decompression (see Table 20-4). The risk of diagnostic EUS is comparable with that of diagnostic upper endoscopy; when needle biopsy is used, the mortality rate is [62] approximately 0.1%. EUS may be most useful in circumstances in which the patient is thought to be at high risk for complications of ERCP or percutaneous THC.

Nuclear Imaging Studies

Nuclear scintigraphy of the biliary tree, although helpful in the diagnosis of cholecystitis, is not sufficiently sensitive to justify its routine use in the diagnostic evaluation of [45,46] jaundice. Furthermore, hepatic uptake of radiolabeled derivatives of iminodiacetic acid [63] (e.g., HIDA) is limited when the serum bilirubin level exceeds 7 to 10 mg/ dL. One exception to this generalization is in the evaluation of a potential bile leak, an uncommon cause of

jaundice following biliary surgery, in which scintigraphy has an accuracy rate as high as [64] 87%.

Suggested Strategies for Imaging

The order of imaging studies depends largely on the clinical likelihood of obstructive jaundice (see Fig. 20-2). Several diagnostic strategies have been compared by clinical decision [65] analysis. If the probability of biliary obstruction is approximately 20%, the positive and negative predictive values of a strategy that uses ultrasonography as the initial test is estimated to be 96% and 98%, respectively. This strategy compares favorably with one that uses ERCP as the initial test. Alternatively, if the probability of biliary obstruction is 60%, a strategy that uses ultrasonography as the first test would yield a positive predictive value of 99%, whereas the negative predictive value would fall to 89%. The implication is that if the level of suspicion for biliary tract obstruction is high and an ultrasound does not show dilated bile ducts, further studies to visualize the biliary tree should be pursued. Therefore, in jaundiced patients in whom biliary obstruction is a possibility, abdominal ultrasonography (or CT) is an appropriate initial approach. If the bile ducts are dilated, the biliary tree should be imaged directly with ERCP (or percutaneous THC) and appropriate therapy undertaken if biliary obstruction is found. If the bile ducts are not dilated on abdominal ultrasonography (or CT), the next step depends on the clinical likelihood of biliary obstruction. If the likelihood of biliary obstruction is thought to be low, the patient should be evaluated for intrinsic liver disease (see later). If the likelihood of biliary obstruction is believed to be intermediate, MRCP or EUS is a reasonable option for imaging the biliary tree before an evaluation for a hepatic disorder is undertaken. Among patients in whom biliary obstruction is believed to be likely, ERCP (or percutaneous THC) should be considered as the next step. If ERCP or percutaneous THC does not show biliary obstruction, the patient should be evaluated for cholestatic liver disease. The decision to use ERCP versus percutaneous THC will be influenced by various factors (see Table 20-4), including the availability of each procedure at a particular facility, presence or absence of dilated bile ducts on initial imaging, and suspected level of biliary obstruction. Under most circumstances, ERCP should be the procedure of choice, because it is comparable with percutaneous THC in accuracy, technical success rate, and frequency of major complications; tends to be more widely available; and may offer better postprocedure tolerability (e.g., no need for an external biliary drainage tube).

OTHER STUDIES Serologic Testing

When imaging studies do not suggest biliary obstruction, jaundiced patients with biochemical evidence of hepatocellular dysfunction or cholestasis should be evaluated for underlying liver disease. Depending on the disorder suspected, screening laboratory studies may include viral serologies (including those for hepatitis B and C and, if the disease is acute, hepatitis A), serum levels of iron, transferrin, and ferritin (for hemochromatosis), ceruloplasmin (for Wilson disease), antimitochondrial antibodies (for primary biliary cirrhosis), antinuclear antibodies, smooth muscle antibodies, and serum protein electrophoresis or serum immunoglobulins (for autoimmune hepatitis), and tissue transglutaminase antibodies (for celiac disease). Confirmation of these diagnoses, as well as elucidation of diagnoses not revealed by serologic analysis, may be made by liver biopsyand small bowel biopsy in the case of celiac disease.

Liver Biopsy
Liver biopsy provides precise information regarding hepatic lobular architecture and extent and pattern of fibrosis, and is most helpful for patients with persistent and undiagnosed jaundice. With special histologic stains and, if appropriate, quantification of copper or iron content, liver biopsy permits the diagnosis of viral hepatitis, fatty liver disease, hemochromatosis, Wilson disease, primary biliary cirrhosis, granulomatous hepatitis, and neoplasms. Occasionally, liver biopsy specimens provide clues to otherwise unsuspected biliary tract obstruction, the histologic features of which are shown inFigure 20-3; however, liver histology may be entirely normal in acute biliary obstruction. Liver biopsy is associated with a low but definite complication rate, predominantly from bleeding and perforation, and the

need for hospitalization in 1% of cases; the mortality rate is approximately 0.01% [66] (see Chapter 73).
Figure 20-3. Liver histology in biliary tract obstruction. A, Prominent bile duct proliferation (arrows) and a mixed portal-based inflammatory infiltrate are evident. Periportal hepatocytes show feathery degeneration (arrowheads), indicative of cholate stasis, cytological changes caused by prolonged cholestasis (Hematoxylin and eosin, 200). B, The periportal bilirubin-stained region (arrow) surrounded by necrotic cells represents a bile infarct (Hematoxylin and eosin, 40).