Tavankit Singh, MD

Liver cirrhosis is implicated in 75% to 85% of ascites cases in the Western world, with heart failure or malignancy accounting for fewer cases.¹ Among patients who have decompensated cirrhosis with ascites, annual mortality is 20%.² Another study showed a 3-year survival rate after onset of ascites of only 56%.³ It is vital for primary care physicians (PCPs) to be alert for ascites not only in patients who have risk factors for chronic liver disease and cirrhosis—eg, a history of alcohol use disorder, chronic viral infections (hepatitis B and C), or metabolic syndrome—but also in patients with abnormal liver function tests and thrombocytopenia. In this review, we discuss the initial assessment of ascites and its long-term management, concentrating on the role of the PCP.

Pathophysiology: Vasodilation leads to a cascade

Splanchnic vasodilation is the main underlying event triggering a pathologic cascade that leads to the development of ascites.⁴ Initially portal hypertension in the setting of liver inflammation and fibrosis causes the release of inflammatory cytokines such as nitric oxide and carbon monoxide. This, in turn, causes the pathologic dilation of splanchnic circulation that decreases effective circulating volume. Activation of the sympathetic nervous system, vasopressin, and renin-­angiotensin-aldosterone system (RAAS) then causes the proximal and distal tubules to increase renal absorption of sodium and water.⁵ The resulting volume overload further decreases the heart’s ability to maintain circulating volume, leading to increased activation of compensating symptoms. This vicious cycle eventually manifests as ascites.⁶

A complex interplay of cirrhosis-associated immune dysfunction (CAID), gut dysbiosis, and increased translocation of microorganisms into ascitic fluid is also an important aspect of the pathogenesis.⁷ CAID (FIGURE 1)^7,8 is an immunodeficient state due to cirrhosis with reduced phagocytic activity by neutrophils and macrophages, T- and B-cell hypoproliferation, and reduced cytotoxicity of natural killer cells. In parallel, there is increased production of inflammatory cytokines due to the effects of damage-associated molecular patterns (DAMPs) from hepatocytes and ­pathogen-associated molecular patterns (PAMPs) from the gut microbiota on the immune system, which leads to many of the manifestations of decompensated cirrhosis including ascites.⁸

Key in on these elementsof the history and exam

Each step of the basic work-up for ascites provides opportunities to refine or redirect the diagnostic inquiry (TABLE).

History

Generally, patients with ascites present with weight gain and symptoms of abdominal distension, such as early satiety, nausea, and vomiting. Besides cirrhosis, rule out other causes of ascites, as treatment differs based on the cause.⁹ Also ask about histories of cancer and cardiac, renal, or thyroid disease.¹⁰

Patients with ascites in the setting of liver disease usually are asymptomatic in its early stages. Common complaints are vague abdominal pain, generalized weakness, malaise, and fatigue.¹¹ Ask patients about risk factors for liver disease such as obesity, diabetes, hypertension, alcohol use, unsafe sexual practices, recent travel, and needle sharing or drug use. Due to a strong association between obstructive sleep apnea and fatty liver disease, consider screening at-risk patients for sleep apnea.¹²

Physical exam

When there are risk factors for liver disease, examine the patient for stigmata of cirrhosis and ascites. Signs of liver disease, aside from ascites, may include spider angiomas on the upper trunk (33% of cirrhosis patients),¹³ gynecomastia (44% of cirrhosis patients),¹⁴ palmar erythema, jaundice, asterixis, and abdominal wall collaterals including caput medusa.¹⁵

Continue to: We suggest a systematic...

We suggest a systematic and targeted approach to using various physical exam maneuvers described in the literature. If the patient has a full/distended abdomen, percuss the flanks. If increased dullness at the flanks is detected, check for shifting dullness, which indicates at least 1500 mL of fluid in the abdomen.¹⁶ Keep in mind that a 10% chance of ascites exists even if shifting dullness is absent.¹⁷ Maneuvers such as the puddle sign and fluid thrill are less accurate than shifting dullness, which has 83% sensitivity and 56% specificity in detecting ascites.¹⁷ Patients with cirrhosis also have a high likelihood of complications from ascites such as inguinal, umbilical, and other hernias.

Diagnostic work-up includes blood tests and ultrasound

Blood tests. The initial work-up for ascites should include complete blood count, complete metabolic panel, and prothrombin time/international normalized ratio.¹⁸

Abdominal ultrasound is recommended as the first-line imaging test.¹⁹ Aside from detecting ascites, it can give an estimate of the volume of ascites and indicate whether it is amenable to paracentesis. A vascular exam added to the standard ultrasound can detect radiologic evidence of portal hypertension such as splenomegaly, portosystemic collaterals, splenorenal shunt, patency of the paraumbilical vein, and portal vein diameter. Patients with established cirrhosis also require abdominal ultrasound every 6 months to screen for hepatocellular cancer.²⁰

Abdominal paracentesis is the cornerstone of ascites evaluation.²¹ It is indicated for every patient with new-onset ascites or for any patient with known ascites and clinical deterioration. Ascitic fluid analysis can be used to easily differentiate portal hypertension from other causes of ascites. It can also be used to rule out bacterial peritonitis. The recommended sites for evaluation are in the left lower quadrant, 3 cm cranially and 3 cm medially from the anterior superior iliac spine.²² A large cohort study showed that abdominal ultrasound-guided paracentesis reduced bleeding complications by 68% following the procedure and is strongly recommended (if available).²³ Generally, paracentesis is a relatively safe procedure with a low risk of complications such as abdominal wall hematoma (1%), hemoperitoneum (< 0.1%), bowel perforation (< 0.1%), and infection (< 0.1%).²⁴

Calculating the serum ascites albumin gradient better characterizes ascitic fluid than total protein-based tests.

Assess all ascitic fluid samples for color, consistency, cell count and differential, albumin, and total protein. These tests are usually sufficient to provide evidence regarding the cause of ascites. If there is suspicion of infection, order a gram stain and culture (80% sensitivity for detecting an infection if obtained prior to initiation of antibiotics)²⁵ and glucose, lactate dehydrogenase (useful to differentiate primary from secondary bacterial peritonitis),²⁶ and amylase tests. Other tests such as cytology, acid-fast bacilli smear and culture, and triglyceride level should only be obtained if specific conditions are suspected based on high pretest probabilities.

Continue to: Calculating serum ascites albumin gradient...

Calculating serum ascites albumin gradient (SAAG) is recommended as it has been shown to better characterize ascitic fluid than total protein-based tests.²⁷ SAAG is calculated by subtracting the level of ascitic fluid albumin from serum albumin level (SAAG = serum albumin – ascitic fluid albumin). A SAAG ≥ 1.1 g/dL is consistent with portal hypertension,²⁸ with approximately 97% accuracy.

After calculating SAAG, look at total protein levels in ascitic fluid. Total protein concentration ≥ 2.5 g/dL with SAAG ≥ 1.1 g/dL has a 78.3% diagnostic accuracy in determining heart failure as the cause of ascites, with a sensitivity of 53.3% and specificity of 86.7%.²⁸ On the other hand, a value of total protein < 2.5 g/dL indicates cirrhosis, liver failure, or acute hepatitis as the cause of fluid build-up.²⁹ Stepwise evaluation of SAAG and total protein and how they can point toward the most likely cause of ascites is presented in FIGURE 2.^27-29

Management

Noninvasive measures

Sodium restriction. The aim of treatment for uncomplicated clinically apparent ascites is sodium restriction and removal of fluid from the body. Dietary salt restriction is complicated, and care should be taken to properly educate patients. Salt restriction advised in the literature has shifted from a strict measure of < 2 g/d³⁰ to more moderate strategies (described below).¹⁸

The 2 main reasons for this easing of restriction are issues with patient compliance and concerns about adverse effects with aggressive salt-restricted diets. One study assessing patient compliance with a salt-restricted diet found that more than two-thirds of the patients were noncompliant,³¹ and 65% of the patients incorrectly assumed they were following the plan, which suggests poor dietary education.³¹ Of the group that was compliant, 20% actually decreased their caloric intake, which can be detrimental in liver disease.³¹ Concerns have been raised that aggressive salt restriction along with diuretic use can lead to diuretic-induced hyponatremia and renal failure.³² Current European Association for the Study of the Liver (EASL) guidelines recommend salt restriction to a more moderate degree (80-120 mmol/d of sodium). This is equivalent to 4.9-6.9 g of salt (1 tablespoon is roughly equivalent to 6 g or 104 mmol of sodium).¹⁸

Diuretics. Initiation and dosage of diuretic therapy is a matter of some controversy. Historically, simultaneous ­administration of a loop diuretic and mineralocorticoid receptor blocker were recommended: 40 mg furosemide and 100 mg spironolactone, keeping the ratio constant with any dosage increases. This was based on a randomized controlled trial (RCT) showing that the combined diuretic therapy effectively mobilized ascites in a shorter period of time and with less frequent adverse effects (eg, hyperkalemia) compared with initial monotherapy.³³

Continue to: On the other hand...

On the other hand, another study with more stable patients and relatively normal renal function showed that starting with a mineralocorticoid receptor blocker alone with sequential dose increments had equivalent benefit with no increase in adverse effects.³⁴ Since the patient population in this study was more in line with what a PCP might encounter, we recommend following this guideline initially and keeping a close watch on serum electrolytes.

Usual maximum doses are spironolactone 400 mg/d and furosemide 160 mg/d.^21,35 Adequate weight loss for patients with diffuse edema is at least 1 kg/d, per EASL guidelines.^36,37 However, this might not be practical in outpatient settings, and a more conservative target of 0.5 kg/d may be used for patients without significant edema.³⁷

It is vital to get accurate daily weights and avoid excessive diuretic use, as it has been associated with intravascular volume depletion and acute kidney injury (25%), hyponatremia (28%),^38,39 and hepatic encephalopathy (30%).⁴⁰ Therefore, patients with acute kidney injury, hyponatremia, acute variceal hemorrhage, or infection should also have their diuretics held until their creatinine returns to baseline.

Invasive measures

Large-volume paracentesis. Patients with extensive and tense ascites should be treated initially with large-volume paracentesis, as this has been shown to predictably remove fluid more effectively than diuretics.³⁸ This should be accompanied by albumin administration, 8 g for every liter of ascitic fluid removed if the total amount exceeds 5 L.⁴¹ Following large-volume paracentesis, manage patients with the standard salt restriction and diuretic regimen.³⁸ Serial large-volume paracentesis is a temporary measure reserved for a select group of patients who are intolerant to diuretics and are not candidates for a shunt.

Transjugular intrahepatic portosystemic shunt (TIPS) is another option to control refractory ascites, but its benefit should be weighed against complications such as hepatic encephalopathy. An RCT found that TIPS with covered stents improved survival in patients with cirrhosis compared with regular large-volume paracentesis.⁴² Patients should be referred to hepatologists to make a determination about TIPS placement. Widely accepted contraindications for the placement of TIPS are decompensated cirrhosis (Child-Pugh > 11, model for end-stage liver disease [MELD] > 18), renal failure (serum creatinine > 3 mg/dL), heart failure, porto-pulmonary hypertension, and uncontrolled sepsis.⁴³ Recurrent or persistent hepatic encephalopathy (West Haven grade ≥ 2) is also a contraindication. The West Haven scale is widely used to measure severity of hepatic encephalopathy, grading it from 1 to 4, with 1 being mild encephalopathy characterized by lack of awareness and shorter attention span, and 4 indicating unresponsiveness or coma.⁴⁴

Continue to: How to manage refractory ascites

Fragile patients are those with refractory ascites that is either unresponsive to standard salt restriction and maximum-dose diuretic therapy or that results in a re-accumulation of ascitic fluid soon after paracentesis.⁴⁵ Specialist care is required to improve survival and quality of life for these patients. They should be referred to a hepatologist for consideration of TIPS placement or liver transplantation.¹⁸

Long-term use of albumin was tested in 2 trials for management of decompensated cirrhosis with ascites, yielding conflicting results. The ANSWER trial from Italy showed benefit with this treatment for prolonged survival.⁴⁶ The other trial, from Spain, showed no benefit from albumin and midodrine administration for survival or for improving complications of cirrhosis.⁴⁷ The contradictory results are likely due to heterogeneous populations in the 2 trials and differences in dose and duration of albumin administration. Hence, no clear recommendations can be made based on the available data; further research is needed.

Getting a handle on bacterial peritonitis

Bacterial peritonitis can be divided into spontaneous bacterial peritonitis (SBP) and secondary bacterial peritonitis. SBP is a common complication in patients with cirrhosis and occurs in around 16% of hospitalized patients, based on 1 study.⁴⁸ SBP is defined as a polymorphonuclear leukocyte count ≥ 250 cells/μL in the absence of a surgically treatable source of infection.⁴⁹ It is believed to be caused by bacterial translocation and is treated empirically with a third-­generation cephalosporin. This treatment has been shown to be effective in 85% of patients.⁵⁰

Diuresis with mineralocorticoid inhibitors alone may be considered for new onset mild-to-moderate ascites in patients with normal renal function.

Patients with SBP are at a higher risk for renal impairment, likely resulting from increased cytokine production and decreased circulatory volume.⁵¹ Concomitant albumin administration has been shown to significantly improve outcomes and to reduce rates of hepatorenal syndrome in patients with serum creatinine > 1 mg/dL, blood urea nitrogen > 30 mg/dL, or total bilirubin > 4 mg/dL.⁵² The recommended amount of albumin is 1.5 g/kg given within 6 hours of SBP detection and repeat administration of 1 g/kg on Day 3.⁵²

Guidelines from the American Association for the Study of Liver Diseases and from EASL recommend the long-term use of daily norfloxacin or trimethoprim-­sulfamethoxazole as secondary prophylaxis in patients who have survived an episode of SBP.^18,30 Long-term antibiotic use is also justified for primary prophylaxis in cirrhosis patients who fulfill certain criteria: ascitic fluid protein < 1.5 g/dL along with impaired renal function (serum creatinine ≥ 1.2 mg/dL, blood urea nitrogen ≥ 25 mg/dL, or serum sodium ≥ 130 mEq/L) or with decompensated cirrhosis (Child-Pugh score ≥ 9 and bilirubin ≥ 3 mg/dL).⁵³ It has been shown to reduce the risk of SBP and hepatorenal syndrome, and improve overall survival.⁵³

Continue to: Avoid these medications

Avoid these medications

Commonly used medications that should be avoided in patients with cirrhosis and ascites are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers. These agents block the action of angiotensin, which is a vital vasoconstrictor, and thereby cause a drop in blood pressure. This has independently been associated with poor outcomes in patients with cirrhosis.³⁷

Commonly used medications that should be avoided in patients with cirrhosis and ascites are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers.

Nonsteroidal anti-inflammatory drugs (NSAIDs) are also relatively contraindicated in cirrhosis, as they can affect kidney function, induce azotemia, and reduce kidney sodium excretion. NSAIDs induce vasoconstriction of afferent arterioles in the kidneys, leading to a decreased glomerular filtration rate, further activating RAAS and sympathetic drive. This leads to increased sodium and water retention and worsening ascites.⁵⁴

Improve outcomes by circling in a hepatologist

PCPs can play a vital role in the prevention, treatment, surveillance, and home care of patients with cirrhosis who are at risk for ascites.⁵⁵ Referral of patients with hepatic impairment manifesting as unexplained abnormal liver function tests, new-onset ascites, and/or image findings consistent with cirrhosis to a hepatologist at least once is recommended. Such referrals have been shown to be associated with a better overall outcome.⁵⁶ Patients with known cirrhosis leading to ascites can generally be managed at home with the assistance of specialists and specialized nurses.³⁵

NSAIDs are relatively contraindicated in cirrhosis as they can affect kidney function, induce azotemia, and reduce kidney sodium excretion.

In a study from the University of Michigan, 69% of patients with cirrhosis had at least 1 nonelective readmission; 14% of patients were readmitted within 1 week, and 37% within 1 month.⁵⁷ These are staggering statistics that highlight the gaps in care coordination and management of patients with cirrhosis in the outpatient setting. PCPs can play a vital role in bridging this gap.

A promising framework is suggested by a study from Italy by Morando et al in 2013.⁵⁸ The researchers assessed a specialized health care model for cirrhotic patients and showed significant improvement in health care cost, readmission rate, and overall mortality when compared with the existing model of outpatient care.⁵⁸

Continue to: This was not a blinded study...

This was not a blinded study and there were concerns raised by the scientific community about its design. Because it was conducted in Italy, the results might not be fully applicable to the United States health care setting. However, it did show that better coordination of care leads to significantly better patient outcomes and reduces health care expenditure. Therefore, a more complete understanding of the disease process and latest literature by PCPs, communication with specialists, and comprehensive coordination of care by all parties involved is vital for the management of this patient population.

CORRESPONDENCE
Muhammad Salman Faisal, MD, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195; faisalm@ccf.org

Liver cirrhosis is implicated in 75% to 85% of ascites cases in the Western world, with heart failure or malignancy accounting for fewer cases.¹ Among patients who have decompensated cirrhosis with ascites, annual mortality is 20%.² Another study showed a 3-year survival rate after onset of ascites of only 56%.³ It is vital for primary care physicians (PCPs) to be alert for ascites not only in patients who have risk factors for chronic liver disease and cirrhosis—eg, a history of alcohol use disorder, chronic viral infections (hepatitis B and C), or metabolic syndrome—but also in patients with abnormal liver function tests and thrombocytopenia. In this review, we discuss the initial assessment of ascites and its long-term management, concentrating on the role of the PCP.

Pathophysiology: Vasodilation leads to a cascade

Splanchnic vasodilation is the main underlying event triggering a pathologic cascade that leads to the development of ascites.⁴ Initially portal hypertension in the setting of liver inflammation and fibrosis causes the release of inflammatory cytokines such as nitric oxide and carbon monoxide. This, in turn, causes the pathologic dilation of splanchnic circulation that decreases effective circulating volume. Activation of the sympathetic nervous system, vasopressin, and renin-­angiotensin-aldosterone system (RAAS) then causes the proximal and distal tubules to increase renal absorption of sodium and water.⁵ The resulting volume overload further decreases the heart’s ability to maintain circulating volume, leading to increased activation of compensating symptoms. This vicious cycle eventually manifests as ascites.⁶

A complex interplay of cirrhosis-associated immune dysfunction (CAID), gut dysbiosis, and increased translocation of microorganisms into ascitic fluid is also an important aspect of the pathogenesis.⁷ CAID (FIGURE 1)^7,8 is an immunodeficient state due to cirrhosis with reduced phagocytic activity by neutrophils and macrophages, T- and B-cell hypoproliferation, and reduced cytotoxicity of natural killer cells. In parallel, there is increased production of inflammatory cytokines due to the effects of damage-associated molecular patterns (DAMPs) from hepatocytes and ­pathogen-associated molecular patterns (PAMPs) from the gut microbiota on the immune system, which leads to many of the manifestations of decompensated cirrhosis including ascites.⁸

Key in on these elementsof the history and exam

Each step of the basic work-up for ascites provides opportunities to refine or redirect the diagnostic inquiry (TABLE).

History

Generally, patients with ascites present with weight gain and symptoms of abdominal distension, such as early satiety, nausea, and vomiting. Besides cirrhosis, rule out other causes of ascites, as treatment differs based on the cause.⁹ Also ask about histories of cancer and cardiac, renal, or thyroid disease.¹⁰

Patients with ascites in the setting of liver disease usually are asymptomatic in its early stages. Common complaints are vague abdominal pain, generalized weakness, malaise, and fatigue.¹¹ Ask patients about risk factors for liver disease such as obesity, diabetes, hypertension, alcohol use, unsafe sexual practices, recent travel, and needle sharing or drug use. Due to a strong association between obstructive sleep apnea and fatty liver disease, consider screening at-risk patients for sleep apnea.¹²

Physical exam

When there are risk factors for liver disease, examine the patient for stigmata of cirrhosis and ascites. Signs of liver disease, aside from ascites, may include spider angiomas on the upper trunk (33% of cirrhosis patients),¹³ gynecomastia (44% of cirrhosis patients),¹⁴ palmar erythema, jaundice, asterixis, and abdominal wall collaterals including caput medusa.¹⁵

Continue to: We suggest a systematic...

We suggest a systematic and targeted approach to using various physical exam maneuvers described in the literature. If the patient has a full/distended abdomen, percuss the flanks. If increased dullness at the flanks is detected, check for shifting dullness, which indicates at least 1500 mL of fluid in the abdomen.¹⁶ Keep in mind that a 10% chance of ascites exists even if shifting dullness is absent.¹⁷ Maneuvers such as the puddle sign and fluid thrill are less accurate than shifting dullness, which has 83% sensitivity and 56% specificity in detecting ascites.¹⁷ Patients with cirrhosis also have a high likelihood of complications from ascites such as inguinal, umbilical, and other hernias.

Diagnostic work-up includes blood tests and ultrasound

Blood tests. The initial work-up for ascites should include complete blood count, complete metabolic panel, and prothrombin time/international normalized ratio.¹⁸

Abdominal ultrasound is recommended as the first-line imaging test.¹⁹ Aside from detecting ascites, it can give an estimate of the volume of ascites and indicate whether it is amenable to paracentesis. A vascular exam added to the standard ultrasound can detect radiologic evidence of portal hypertension such as splenomegaly, portosystemic collaterals, splenorenal shunt, patency of the paraumbilical vein, and portal vein diameter. Patients with established cirrhosis also require abdominal ultrasound every 6 months to screen for hepatocellular cancer.²⁰

Abdominal paracentesis is the cornerstone of ascites evaluation.²¹ It is indicated for every patient with new-onset ascites or for any patient with known ascites and clinical deterioration. Ascitic fluid analysis can be used to easily differentiate portal hypertension from other causes of ascites. It can also be used to rule out bacterial peritonitis. The recommended sites for evaluation are in the left lower quadrant, 3 cm cranially and 3 cm medially from the anterior superior iliac spine.²² A large cohort study showed that abdominal ultrasound-guided paracentesis reduced bleeding complications by 68% following the procedure and is strongly recommended (if available).²³ Generally, paracentesis is a relatively safe procedure with a low risk of complications such as abdominal wall hematoma (1%), hemoperitoneum (< 0.1%), bowel perforation (< 0.1%), and infection (< 0.1%).²⁴

Calculating the serum ascites albumin gradient better characterizes ascitic fluid than total protein-based tests.

Assess all ascitic fluid samples for color, consistency, cell count and differential, albumin, and total protein. These tests are usually sufficient to provide evidence regarding the cause of ascites. If there is suspicion of infection, order a gram stain and culture (80% sensitivity for detecting an infection if obtained prior to initiation of antibiotics)²⁵ and glucose, lactate dehydrogenase (useful to differentiate primary from secondary bacterial peritonitis),²⁶ and amylase tests. Other tests such as cytology, acid-fast bacilli smear and culture, and triglyceride level should only be obtained if specific conditions are suspected based on high pretest probabilities.

Continue to: Calculating serum ascites albumin gradient...

Calculating serum ascites albumin gradient (SAAG) is recommended as it has been shown to better characterize ascitic fluid than total protein-based tests.²⁷ SAAG is calculated by subtracting the level of ascitic fluid albumin from serum albumin level (SAAG = serum albumin – ascitic fluid albumin). A SAAG ≥ 1.1 g/dL is consistent with portal hypertension,²⁸ with approximately 97% accuracy.

After calculating SAAG, look at total protein levels in ascitic fluid. Total protein concentration ≥ 2.5 g/dL with SAAG ≥ 1.1 g/dL has a 78.3% diagnostic accuracy in determining heart failure as the cause of ascites, with a sensitivity of 53.3% and specificity of 86.7%.²⁸ On the other hand, a value of total protein < 2.5 g/dL indicates cirrhosis, liver failure, or acute hepatitis as the cause of fluid build-up.²⁹ Stepwise evaluation of SAAG and total protein and how they can point toward the most likely cause of ascites is presented in FIGURE 2.^27-29

Management

Noninvasive measures

Sodium restriction. The aim of treatment for uncomplicated clinically apparent ascites is sodium restriction and removal of fluid from the body. Dietary salt restriction is complicated, and care should be taken to properly educate patients. Salt restriction advised in the literature has shifted from a strict measure of < 2 g/d³⁰ to more moderate strategies (described below).¹⁸

The 2 main reasons for this easing of restriction are issues with patient compliance and concerns about adverse effects with aggressive salt-restricted diets. One study assessing patient compliance with a salt-restricted diet found that more than two-thirds of the patients were noncompliant,³¹ and 65% of the patients incorrectly assumed they were following the plan, which suggests poor dietary education.³¹ Of the group that was compliant, 20% actually decreased their caloric intake, which can be detrimental in liver disease.³¹ Concerns have been raised that aggressive salt restriction along with diuretic use can lead to diuretic-induced hyponatremia and renal failure.³² Current European Association for the Study of the Liver (EASL) guidelines recommend salt restriction to a more moderate degree (80-120 mmol/d of sodium). This is equivalent to 4.9-6.9 g of salt (1 tablespoon is roughly equivalent to 6 g or 104 mmol of sodium).¹⁸

Diuretics. Initiation and dosage of diuretic therapy is a matter of some controversy. Historically, simultaneous ­administration of a loop diuretic and mineralocorticoid receptor blocker were recommended: 40 mg furosemide and 100 mg spironolactone, keeping the ratio constant with any dosage increases. This was based on a randomized controlled trial (RCT) showing that the combined diuretic therapy effectively mobilized ascites in a shorter period of time and with less frequent adverse effects (eg, hyperkalemia) compared with initial monotherapy.³³

Continue to: On the other hand...

On the other hand, another study with more stable patients and relatively normal renal function showed that starting with a mineralocorticoid receptor blocker alone with sequential dose increments had equivalent benefit with no increase in adverse effects.³⁴ Since the patient population in this study was more in line with what a PCP might encounter, we recommend following this guideline initially and keeping a close watch on serum electrolytes.

Usual maximum doses are spironolactone 400 mg/d and furosemide 160 mg/d.^21,35 Adequate weight loss for patients with diffuse edema is at least 1 kg/d, per EASL guidelines.^36,37 However, this might not be practical in outpatient settings, and a more conservative target of 0.5 kg/d may be used for patients without significant edema.³⁷

It is vital to get accurate daily weights and avoid excessive diuretic use, as it has been associated with intravascular volume depletion and acute kidney injury (25%), hyponatremia (28%),^38,39 and hepatic encephalopathy (30%).⁴⁰ Therefore, patients with acute kidney injury, hyponatremia, acute variceal hemorrhage, or infection should also have their diuretics held until their creatinine returns to baseline.

Invasive measures

Large-volume paracentesis. Patients with extensive and tense ascites should be treated initially with large-volume paracentesis, as this has been shown to predictably remove fluid more effectively than diuretics.³⁸ This should be accompanied by albumin administration, 8 g for every liter of ascitic fluid removed if the total amount exceeds 5 L.⁴¹ Following large-volume paracentesis, manage patients with the standard salt restriction and diuretic regimen.³⁸ Serial large-volume paracentesis is a temporary measure reserved for a select group of patients who are intolerant to diuretics and are not candidates for a shunt.

Transjugular intrahepatic portosystemic shunt (TIPS) is another option to control refractory ascites, but its benefit should be weighed against complications such as hepatic encephalopathy. An RCT found that TIPS with covered stents improved survival in patients with cirrhosis compared with regular large-volume paracentesis.⁴² Patients should be referred to hepatologists to make a determination about TIPS placement. Widely accepted contraindications for the placement of TIPS are decompensated cirrhosis (Child-Pugh > 11, model for end-stage liver disease [MELD] > 18), renal failure (serum creatinine > 3 mg/dL), heart failure, porto-pulmonary hypertension, and uncontrolled sepsis.⁴³ Recurrent or persistent hepatic encephalopathy (West Haven grade ≥ 2) is also a contraindication. The West Haven scale is widely used to measure severity of hepatic encephalopathy, grading it from 1 to 4, with 1 being mild encephalopathy characterized by lack of awareness and shorter attention span, and 4 indicating unresponsiveness or coma.⁴⁴

Continue to: How to manage refractory ascites

Fragile patients are those with refractory ascites that is either unresponsive to standard salt restriction and maximum-dose diuretic therapy or that results in a re-accumulation of ascitic fluid soon after paracentesis.⁴⁵ Specialist care is required to improve survival and quality of life for these patients. They should be referred to a hepatologist for consideration of TIPS placement or liver transplantation.¹⁸

Long-term use of albumin was tested in 2 trials for management of decompensated cirrhosis with ascites, yielding conflicting results. The ANSWER trial from Italy showed benefit with this treatment for prolonged survival.⁴⁶ The other trial, from Spain, showed no benefit from albumin and midodrine administration for survival or for improving complications of cirrhosis.⁴⁷ The contradictory results are likely due to heterogeneous populations in the 2 trials and differences in dose and duration of albumin administration. Hence, no clear recommendations can be made based on the available data; further research is needed.

Getting a handle on bacterial peritonitis

Bacterial peritonitis can be divided into spontaneous bacterial peritonitis (SBP) and secondary bacterial peritonitis. SBP is a common complication in patients with cirrhosis and occurs in around 16% of hospitalized patients, based on 1 study.⁴⁸ SBP is defined as a polymorphonuclear leukocyte count ≥ 250 cells/μL in the absence of a surgically treatable source of infection.⁴⁹ It is believed to be caused by bacterial translocation and is treated empirically with a third-­generation cephalosporin. This treatment has been shown to be effective in 85% of patients.⁵⁰

Diuresis with mineralocorticoid inhibitors alone may be considered for new onset mild-to-moderate ascites in patients with normal renal function.

Patients with SBP are at a higher risk for renal impairment, likely resulting from increased cytokine production and decreased circulatory volume.⁵¹ Concomitant albumin administration has been shown to significantly improve outcomes and to reduce rates of hepatorenal syndrome in patients with serum creatinine > 1 mg/dL, blood urea nitrogen > 30 mg/dL, or total bilirubin > 4 mg/dL.⁵² The recommended amount of albumin is 1.5 g/kg given within 6 hours of SBP detection and repeat administration of 1 g/kg on Day 3.⁵²

Guidelines from the American Association for the Study of Liver Diseases and from EASL recommend the long-term use of daily norfloxacin or trimethoprim-­sulfamethoxazole as secondary prophylaxis in patients who have survived an episode of SBP.^18,30 Long-term antibiotic use is also justified for primary prophylaxis in cirrhosis patients who fulfill certain criteria: ascitic fluid protein < 1.5 g/dL along with impaired renal function (serum creatinine ≥ 1.2 mg/dL, blood urea nitrogen ≥ 25 mg/dL, or serum sodium ≥ 130 mEq/L) or with decompensated cirrhosis (Child-Pugh score ≥ 9 and bilirubin ≥ 3 mg/dL).⁵³ It has been shown to reduce the risk of SBP and hepatorenal syndrome, and improve overall survival.⁵³

Continue to: Avoid these medications

Avoid these medications

Commonly used medications that should be avoided in patients with cirrhosis and ascites are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers. These agents block the action of angiotensin, which is a vital vasoconstrictor, and thereby cause a drop in blood pressure. This has independently been associated with poor outcomes in patients with cirrhosis.³⁷

Commonly used medications that should be avoided in patients with cirrhosis and ascites are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers.

Nonsteroidal anti-inflammatory drugs (NSAIDs) are also relatively contraindicated in cirrhosis, as they can affect kidney function, induce azotemia, and reduce kidney sodium excretion. NSAIDs induce vasoconstriction of afferent arterioles in the kidneys, leading to a decreased glomerular filtration rate, further activating RAAS and sympathetic drive. This leads to increased sodium and water retention and worsening ascites.⁵⁴

Improve outcomes by circling in a hepatologist

PCPs can play a vital role in the prevention, treatment, surveillance, and home care of patients with cirrhosis who are at risk for ascites.⁵⁵ Referral of patients with hepatic impairment manifesting as unexplained abnormal liver function tests, new-onset ascites, and/or image findings consistent with cirrhosis to a hepatologist at least once is recommended. Such referrals have been shown to be associated with a better overall outcome.⁵⁶ Patients with known cirrhosis leading to ascites can generally be managed at home with the assistance of specialists and specialized nurses.³⁵

NSAIDs are relatively contraindicated in cirrhosis as they can affect kidney function, induce azotemia, and reduce kidney sodium excretion.

In a study from the University of Michigan, 69% of patients with cirrhosis had at least 1 nonelective readmission; 14% of patients were readmitted within 1 week, and 37% within 1 month.⁵⁷ These are staggering statistics that highlight the gaps in care coordination and management of patients with cirrhosis in the outpatient setting. PCPs can play a vital role in bridging this gap.

A promising framework is suggested by a study from Italy by Morando et al in 2013.⁵⁸ The researchers assessed a specialized health care model for cirrhotic patients and showed significant improvement in health care cost, readmission rate, and overall mortality when compared with the existing model of outpatient care.⁵⁸

Continue to: This was not a blinded study...

This was not a blinded study and there were concerns raised by the scientific community about its design. Because it was conducted in Italy, the results might not be fully applicable to the United States health care setting. However, it did show that better coordination of care leads to significantly better patient outcomes and reduces health care expenditure. Therefore, a more complete understanding of the disease process and latest literature by PCPs, communication with specialists, and comprehensive coordination of care by all parties involved is vital for the management of this patient population.

CORRESPONDENCE
Muhammad Salman Faisal, MD, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195; faisalm@ccf.org

Liver cirrhosis is implicated in 75% to 85% of ascites cases in the Western world, with heart failure or malignancy accounting for fewer cases.¹ Among patients who have decompensated cirrhosis with ascites, annual mortality is 20%.² Another study showed a 3-year survival rate after onset of ascites of only 56%.³ It is vital for primary care physicians (PCPs) to be alert for ascites not only in patients who have risk factors for chronic liver disease and cirrhosis—eg, a history of alcohol use disorder, chronic viral infections (hepatitis B and C), or metabolic syndrome—but also in patients with abnormal liver function tests and thrombocytopenia. In this review, we discuss the initial assessment of ascites and its long-term management, concentrating on the role of the PCP.

Pathophysiology: Vasodilation leads to a cascade

Splanchnic vasodilation is the main underlying event triggering a pathologic cascade that leads to the development of ascites.⁴ Initially portal hypertension in the setting of liver inflammation and fibrosis causes the release of inflammatory cytokines such as nitric oxide and carbon monoxide. This, in turn, causes the pathologic dilation of splanchnic circulation that decreases effective circulating volume. Activation of the sympathetic nervous system, vasopressin, and renin-­angiotensin-aldosterone system (RAAS) then causes the proximal and distal tubules to increase renal absorption of sodium and water.⁵ The resulting volume overload further decreases the heart’s ability to maintain circulating volume, leading to increased activation of compensating symptoms. This vicious cycle eventually manifests as ascites.⁶

A complex interplay of cirrhosis-associated immune dysfunction (CAID), gut dysbiosis, and increased translocation of microorganisms into ascitic fluid is also an important aspect of the pathogenesis.⁷ CAID (FIGURE 1)^7,8 is an immunodeficient state due to cirrhosis with reduced phagocytic activity by neutrophils and macrophages, T- and B-cell hypoproliferation, and reduced cytotoxicity of natural killer cells. In parallel, there is increased production of inflammatory cytokines due to the effects of damage-associated molecular patterns (DAMPs) from hepatocytes and ­pathogen-associated molecular patterns (PAMPs) from the gut microbiota on the immune system, which leads to many of the manifestations of decompensated cirrhosis including ascites.⁸

Key in on these elementsof the history and exam

Each step of the basic work-up for ascites provides opportunities to refine or redirect the diagnostic inquiry (TABLE).

History

Generally, patients with ascites present with weight gain and symptoms of abdominal distension, such as early satiety, nausea, and vomiting. Besides cirrhosis, rule out other causes of ascites, as treatment differs based on the cause.⁹ Also ask about histories of cancer and cardiac, renal, or thyroid disease.¹⁰

Patients with ascites in the setting of liver disease usually are asymptomatic in its early stages. Common complaints are vague abdominal pain, generalized weakness, malaise, and fatigue.¹¹ Ask patients about risk factors for liver disease such as obesity, diabetes, hypertension, alcohol use, unsafe sexual practices, recent travel, and needle sharing or drug use. Due to a strong association between obstructive sleep apnea and fatty liver disease, consider screening at-risk patients for sleep apnea.¹²

Physical exam

When there are risk factors for liver disease, examine the patient for stigmata of cirrhosis and ascites. Signs of liver disease, aside from ascites, may include spider angiomas on the upper trunk (33% of cirrhosis patients),¹³ gynecomastia (44% of cirrhosis patients),¹⁴ palmar erythema, jaundice, asterixis, and abdominal wall collaterals including caput medusa.¹⁵

Continue to: We suggest a systematic...

We suggest a systematic and targeted approach to using various physical exam maneuvers described in the literature. If the patient has a full/distended abdomen, percuss the flanks. If increased dullness at the flanks is detected, check for shifting dullness, which indicates at least 1500 mL of fluid in the abdomen.¹⁶ Keep in mind that a 10% chance of ascites exists even if shifting dullness is absent.¹⁷ Maneuvers such as the puddle sign and fluid thrill are less accurate than shifting dullness, which has 83% sensitivity and 56% specificity in detecting ascites.¹⁷ Patients with cirrhosis also have a high likelihood of complications from ascites such as inguinal, umbilical, and other hernias.

Diagnostic work-up includes blood tests and ultrasound

Blood tests. The initial work-up for ascites should include complete blood count, complete metabolic panel, and prothrombin time/international normalized ratio.¹⁸

Abdominal ultrasound is recommended as the first-line imaging test.¹⁹ Aside from detecting ascites, it can give an estimate of the volume of ascites and indicate whether it is amenable to paracentesis. A vascular exam added to the standard ultrasound can detect radiologic evidence of portal hypertension such as splenomegaly, portosystemic collaterals, splenorenal shunt, patency of the paraumbilical vein, and portal vein diameter. Patients with established cirrhosis also require abdominal ultrasound every 6 months to screen for hepatocellular cancer.²⁰

Abdominal paracentesis is the cornerstone of ascites evaluation.²¹ It is indicated for every patient with new-onset ascites or for any patient with known ascites and clinical deterioration. Ascitic fluid analysis can be used to easily differentiate portal hypertension from other causes of ascites. It can also be used to rule out bacterial peritonitis. The recommended sites for evaluation are in the left lower quadrant, 3 cm cranially and 3 cm medially from the anterior superior iliac spine.²² A large cohort study showed that abdominal ultrasound-guided paracentesis reduced bleeding complications by 68% following the procedure and is strongly recommended (if available).²³ Generally, paracentesis is a relatively safe procedure with a low risk of complications such as abdominal wall hematoma (1%), hemoperitoneum (< 0.1%), bowel perforation (< 0.1%), and infection (< 0.1%).²⁴

Calculating the serum ascites albumin gradient better characterizes ascitic fluid than total protein-based tests.

Assess all ascitic fluid samples for color, consistency, cell count and differential, albumin, and total protein. These tests are usually sufficient to provide evidence regarding the cause of ascites. If there is suspicion of infection, order a gram stain and culture (80% sensitivity for detecting an infection if obtained prior to initiation of antibiotics)²⁵ and glucose, lactate dehydrogenase (useful to differentiate primary from secondary bacterial peritonitis),²⁶ and amylase tests. Other tests such as cytology, acid-fast bacilli smear and culture, and triglyceride level should only be obtained if specific conditions are suspected based on high pretest probabilities.

Continue to: Calculating serum ascites albumin gradient...

Calculating serum ascites albumin gradient (SAAG) is recommended as it has been shown to better characterize ascitic fluid than total protein-based tests.²⁷ SAAG is calculated by subtracting the level of ascitic fluid albumin from serum albumin level (SAAG = serum albumin – ascitic fluid albumin). A SAAG ≥ 1.1 g/dL is consistent with portal hypertension,²⁸ with approximately 97% accuracy.

After calculating SAAG, look at total protein levels in ascitic fluid. Total protein concentration ≥ 2.5 g/dL with SAAG ≥ 1.1 g/dL has a 78.3% diagnostic accuracy in determining heart failure as the cause of ascites, with a sensitivity of 53.3% and specificity of 86.7%.²⁸ On the other hand, a value of total protein < 2.5 g/dL indicates cirrhosis, liver failure, or acute hepatitis as the cause of fluid build-up.²⁹ Stepwise evaluation of SAAG and total protein and how they can point toward the most likely cause of ascites is presented in FIGURE 2.^27-29

Management

Noninvasive measures

Sodium restriction. The aim of treatment for uncomplicated clinically apparent ascites is sodium restriction and removal of fluid from the body. Dietary salt restriction is complicated, and care should be taken to properly educate patients. Salt restriction advised in the literature has shifted from a strict measure of < 2 g/d³⁰ to more moderate strategies (described below).¹⁸

The 2 main reasons for this easing of restriction are issues with patient compliance and concerns about adverse effects with aggressive salt-restricted diets. One study assessing patient compliance with a salt-restricted diet found that more than two-thirds of the patients were noncompliant,³¹ and 65% of the patients incorrectly assumed they were following the plan, which suggests poor dietary education.³¹ Of the group that was compliant, 20% actually decreased their caloric intake, which can be detrimental in liver disease.³¹ Concerns have been raised that aggressive salt restriction along with diuretic use can lead to diuretic-induced hyponatremia and renal failure.³² Current European Association for the Study of the Liver (EASL) guidelines recommend salt restriction to a more moderate degree (80-120 mmol/d of sodium). This is equivalent to 4.9-6.9 g of salt (1 tablespoon is roughly equivalent to 6 g or 104 mmol of sodium).¹⁸

Diuretics. Initiation and dosage of diuretic therapy is a matter of some controversy. Historically, simultaneous ­administration of a loop diuretic and mineralocorticoid receptor blocker were recommended: 40 mg furosemide and 100 mg spironolactone, keeping the ratio constant with any dosage increases. This was based on a randomized controlled trial (RCT) showing that the combined diuretic therapy effectively mobilized ascites in a shorter period of time and with less frequent adverse effects (eg, hyperkalemia) compared with initial monotherapy.³³

Continue to: On the other hand...

On the other hand, another study with more stable patients and relatively normal renal function showed that starting with a mineralocorticoid receptor blocker alone with sequential dose increments had equivalent benefit with no increase in adverse effects.³⁴ Since the patient population in this study was more in line with what a PCP might encounter, we recommend following this guideline initially and keeping a close watch on serum electrolytes.

Usual maximum doses are spironolactone 400 mg/d and furosemide 160 mg/d.^21,35 Adequate weight loss for patients with diffuse edema is at least 1 kg/d, per EASL guidelines.^36,37 However, this might not be practical in outpatient settings, and a more conservative target of 0.5 kg/d may be used for patients without significant edema.³⁷

It is vital to get accurate daily weights and avoid excessive diuretic use, as it has been associated with intravascular volume depletion and acute kidney injury (25%), hyponatremia (28%),^38,39 and hepatic encephalopathy (30%).⁴⁰ Therefore, patients with acute kidney injury, hyponatremia, acute variceal hemorrhage, or infection should also have their diuretics held until their creatinine returns to baseline.

Invasive measures

Large-volume paracentesis. Patients with extensive and tense ascites should be treated initially with large-volume paracentesis, as this has been shown to predictably remove fluid more effectively than diuretics.³⁸ This should be accompanied by albumin administration, 8 g for every liter of ascitic fluid removed if the total amount exceeds 5 L.⁴¹ Following large-volume paracentesis, manage patients with the standard salt restriction and diuretic regimen.³⁸ Serial large-volume paracentesis is a temporary measure reserved for a select group of patients who are intolerant to diuretics and are not candidates for a shunt.

Transjugular intrahepatic portosystemic shunt (TIPS) is another option to control refractory ascites, but its benefit should be weighed against complications such as hepatic encephalopathy. An RCT found that TIPS with covered stents improved survival in patients with cirrhosis compared with regular large-volume paracentesis.⁴² Patients should be referred to hepatologists to make a determination about TIPS placement. Widely accepted contraindications for the placement of TIPS are decompensated cirrhosis (Child-Pugh > 11, model for end-stage liver disease [MELD] > 18), renal failure (serum creatinine > 3 mg/dL), heart failure, porto-pulmonary hypertension, and uncontrolled sepsis.⁴³ Recurrent or persistent hepatic encephalopathy (West Haven grade ≥ 2) is also a contraindication. The West Haven scale is widely used to measure severity of hepatic encephalopathy, grading it from 1 to 4, with 1 being mild encephalopathy characterized by lack of awareness and shorter attention span, and 4 indicating unresponsiveness or coma.⁴⁴

Continue to: How to manage refractory ascites

Fragile patients are those with refractory ascites that is either unresponsive to standard salt restriction and maximum-dose diuretic therapy or that results in a re-accumulation of ascitic fluid soon after paracentesis.⁴⁵ Specialist care is required to improve survival and quality of life for these patients. They should be referred to a hepatologist for consideration of TIPS placement or liver transplantation.¹⁸

Long-term use of albumin was tested in 2 trials for management of decompensated cirrhosis with ascites, yielding conflicting results. The ANSWER trial from Italy showed benefit with this treatment for prolonged survival.⁴⁶ The other trial, from Spain, showed no benefit from albumin and midodrine administration for survival or for improving complications of cirrhosis.⁴⁷ The contradictory results are likely due to heterogeneous populations in the 2 trials and differences in dose and duration of albumin administration. Hence, no clear recommendations can be made based on the available data; further research is needed.

Getting a handle on bacterial peritonitis

Bacterial peritonitis can be divided into spontaneous bacterial peritonitis (SBP) and secondary bacterial peritonitis. SBP is a common complication in patients with cirrhosis and occurs in around 16% of hospitalized patients, based on 1 study.⁴⁸ SBP is defined as a polymorphonuclear leukocyte count ≥ 250 cells/μL in the absence of a surgically treatable source of infection.⁴⁹ It is believed to be caused by bacterial translocation and is treated empirically with a third-­generation cephalosporin. This treatment has been shown to be effective in 85% of patients.⁵⁰

Diuresis with mineralocorticoid inhibitors alone may be considered for new onset mild-to-moderate ascites in patients with normal renal function.

Patients with SBP are at a higher risk for renal impairment, likely resulting from increased cytokine production and decreased circulatory volume.⁵¹ Concomitant albumin administration has been shown to significantly improve outcomes and to reduce rates of hepatorenal syndrome in patients with serum creatinine > 1 mg/dL, blood urea nitrogen > 30 mg/dL, or total bilirubin > 4 mg/dL.⁵² The recommended amount of albumin is 1.5 g/kg given within 6 hours of SBP detection and repeat administration of 1 g/kg on Day 3.⁵²

Guidelines from the American Association for the Study of Liver Diseases and from EASL recommend the long-term use of daily norfloxacin or trimethoprim-­sulfamethoxazole as secondary prophylaxis in patients who have survived an episode of SBP.^18,30 Long-term antibiotic use is also justified for primary prophylaxis in cirrhosis patients who fulfill certain criteria: ascitic fluid protein < 1.5 g/dL along with impaired renal function (serum creatinine ≥ 1.2 mg/dL, blood urea nitrogen ≥ 25 mg/dL, or serum sodium ≥ 130 mEq/L) or with decompensated cirrhosis (Child-Pugh score ≥ 9 and bilirubin ≥ 3 mg/dL).⁵³ It has been shown to reduce the risk of SBP and hepatorenal syndrome, and improve overall survival.⁵³

Continue to: Avoid these medications

Avoid these medications

Commonly used medications that should be avoided in patients with cirrhosis and ascites are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers. These agents block the action of angiotensin, which is a vital vasoconstrictor, and thereby cause a drop in blood pressure. This has independently been associated with poor outcomes in patients with cirrhosis.³⁷

Commonly used medications that should be avoided in patients with cirrhosis and ascites are angiotensin-converting enzyme inhibitors and angiotensin receptor blockers.

Nonsteroidal anti-inflammatory drugs (NSAIDs) are also relatively contraindicated in cirrhosis, as they can affect kidney function, induce azotemia, and reduce kidney sodium excretion. NSAIDs induce vasoconstriction of afferent arterioles in the kidneys, leading to a decreased glomerular filtration rate, further activating RAAS and sympathetic drive. This leads to increased sodium and water retention and worsening ascites.⁵⁴

Improve outcomes by circling in a hepatologist

PCPs can play a vital role in the prevention, treatment, surveillance, and home care of patients with cirrhosis who are at risk for ascites.⁵⁵ Referral of patients with hepatic impairment manifesting as unexplained abnormal liver function tests, new-onset ascites, and/or image findings consistent with cirrhosis to a hepatologist at least once is recommended. Such referrals have been shown to be associated with a better overall outcome.⁵⁶ Patients with known cirrhosis leading to ascites can generally be managed at home with the assistance of specialists and specialized nurses.³⁵

NSAIDs are relatively contraindicated in cirrhosis as they can affect kidney function, induce azotemia, and reduce kidney sodium excretion.

In a study from the University of Michigan, 69% of patients with cirrhosis had at least 1 nonelective readmission; 14% of patients were readmitted within 1 week, and 37% within 1 month.⁵⁷ These are staggering statistics that highlight the gaps in care coordination and management of patients with cirrhosis in the outpatient setting. PCPs can play a vital role in bridging this gap.

A promising framework is suggested by a study from Italy by Morando et al in 2013.⁵⁸ The researchers assessed a specialized health care model for cirrhotic patients and showed significant improvement in health care cost, readmission rate, and overall mortality when compared with the existing model of outpatient care.⁵⁸

Continue to: This was not a blinded study...

This was not a blinded study and there were concerns raised by the scientific community about its design. Because it was conducted in Italy, the results might not be fully applicable to the United States health care setting. However, it did show that better coordination of care leads to significantly better patient outcomes and reduces health care expenditure. Therefore, a more complete understanding of the disease process and latest literature by PCPs, communication with specialists, and comprehensive coordination of care by all parties involved is vital for the management of this patient population.

CORRESPONDENCE
Muhammad Salman Faisal, MD, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195; faisalm@ccf.org

All cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus.¹ But most cases of Barrett esophagus go undiagnosed. And Barrett esophagus is seen in 5% to 15% of patients with gastroesophageal reflux disease.² These facts clearly emphasize the need for screening. Here, we review the rationale and recommendations for screening and surveillance, as well as the range of treatment options.

SCOPE OF THE PROBLEM

The American Cancer Society estimated there were 17,290 new cases of esophageal cancer and 15,850 deaths from it in the United States in 2018.³ Of the 2 main histologic types of esophageal cancer, adenocarcinoma and squamous cell cancer, adenocarcinoma is more common in the United States.

The precursor lesion is Barrett esophagus, defined as an extension of salmon-colored mucosa at least 1 cm into the tubular esophagus proximal to the gastroesophageal junction, with biopsy confirmation of intestinal metaplasia.⁴

The natural course of progression to dysplasia and cancer in Barrett esophagus is unknown but is thought to be stepwise, from no dysplasia to low-grade dysplasia to high-grade dysplasia and cancer, and the cancer risk depends on the degree of dysplasia: the annual risk is 0.33% if there is no dysplasia, 0.54% with low-grade dysplasia, and 7% with high-grade dysplasia.⁴

Although all cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus,¹ more than 90% of patients with newly diagnosed esophageal adenocarcinoma do not have a prior diagnosis of Barrett esophagus.⁵ Therefore, there is a substantial unmet need to expand screening for Barrett esophagus in people at risk.

GASTROESOPHAGEAL REFLUX DISEASE IS A RISK FACTOR FOR CANCER

The rationale behind screening is that detecting Barrett esophagus early and intervening in a timely manner in patients at higher risk of developing adenocarcinoma will decrease mortality.

Chronic gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma (odds ratio [OR] 7.7, 95% confidence interval [CI] 5.3–11.4), and the risk increases when symptoms are long-standing (> 20 years) or severe (OR 43.5, 95% CI 18.3–103.5) or occur daily (OR 5.5, 95% CI 3.2–9.3).⁶

Reflux symptoms are scored as follows:

• Heartburn only, 1 point
• Regurgitation only, 1 point
• Heartburn with regurgitation, 1.5 points
• Nightly symptoms (2 points if yes, 0 if no)
• Symptoms once a week, 0 points; 2 to 6 times a week, 1 point; 7 to 15 times a week, 2 points; more than 15 times a week, 3 points.⁶

A score of 4.5 or higher indicates severe reflux disease. However, it is worth noting that the annual incidence of esophageal adenocarcinoma in patients with long-term gastroesophageal reflux disease is less than 0.001%.⁷

RISK FACTORS FOR BARRETT ESOPHAGUS

Risk factors for Barrett esophagus include:

Male sex. Barrett esophagus is more prevalent in men than in women, at a ratio of 2 to 1; but in individuals under age 50, the ratio is 4 to 1.⁸

Age 50 or older. Barrett esophagus is usually diagnosed in the sixth to seventh decade of life, and the prevalence increases from 2.1% in the third decade to 9.3% in the sixth decade.⁹

White race. It is more prevalent in whites than in blacks (5.0% vs 1.5%, P < .0001).¹⁰

Central obesity. Waist circumference is an independent risk factor: every 5-cm increase carries an OR of 1.14 (95% CI 1.03–1.27, P = .02).¹¹

Cigarette smoking increases the risk of Barrett esophagus (OR 1.42; 95% CI 1.15–1.76).¹²

A family history of Barrett esophagus or esophageal adenocarcinoma is a strong risk factor (OR 12, 95% CI 3.3–44.8). In 1 study, the risk in first- and second-degree relatives of patients with Barrett esophagus was 24%, compared with 5% in a control population (P < .005).¹³

SCREENING GUIDELINES AND DRAWBACKS

The standard screening method is esophagogastroduodenoscopy with sedation, with careful visual inspection and 4-quadrant biopsies every 2 cm using the Seattle protocol, ie, including biopsy of any mucosal irregularities in salmon-colored mucosa above the gastroesophageal junction (Figure 1).⁴

Endoscopic screening is cost-effective, costing $10,440 per quality-adjusted life-year saved, which is well below the accepted threshold of less than $100,000.¹⁴ However, it is still expensive, invasive, and not ideal for screening large populations.

Less-invasive methods under study

Less-invasive, less-expensive methods being tested for mass screening include:

Unsedated transnasal endoscopy. Done with only topical anesthesia, it has high diagnostic accuracy and is quicker and more cost-effective than standard esophagogastroduodenoscopy, with fewer adverse effects. However, the procedure has not yet gained widespread acceptance for regular use by gastroenterologists.¹⁵

A swallowable sponge. Another promising test is cell collection using the Cytosponge Cell Collection Device (Medtronic, Minneapolis, MN). An encapsulated compressed sponge with a string attached is swallowed; in the stomach, the capsule dissolves, and the sponge expands and is then withdrawn using the attached string. The obtained cytology sample from the lower esophagus is then tested for trefoil factor 3, a protein biomarker for Barrett esophagus.¹⁶

A retractable balloon. The EsoCheck Cell Collection Device is a retractable balloon attached to a string. When swallowed, it gathers distal esophageal cells for detecting methylated DNA markers for Barrett esophagus.¹⁷

Esophageal capsule endoscopy uses a camera to visualize the esophagus, but lacks the ability to obtain biopsy samples.

Other screening methods are being tested, although data are limited. Liquid biopsy uses a blood sample to detect microRNAs that are dysregulated in cancer. The “electronic nose” is a device that detects exhaled volatile organic compounds altered in Barrett esophagus. Another test involves taking an oral wash sample to study the oral microbiome for a pattern associated with adenocarcinoma.^18–21

Surveillance in Barrett esophagus aims to detect premalignant changes or early-stage adenocarcinoma to provide longer survival and lower cancer-related mortality. Recent evidence suggests that patients with esophageal adenocarcinoma that is diagnosed in a Barrett esophagus surveillance program have an earlier stage of disease and therefore a survival benefit.²²

Patient education is essential

Before enrolling a patient in a surveillance program, the clinician should explain the risks, benefits, and limitations, the importance of periodic endoscopy, and the possible eventual need for endoscopic therapy or surgery.

The endoscopic procedure

Surveillance involves examination by high-definition white-light endoscopy, with random 4-quadrant biopsies every 2 cm (or every 1 cm in patients with a history of dysplasia) and biopsy of any mucosal irregularity (nodule, ulcer, or other visible lesion). The degree of dysplasia determines the frequency of follow-up surveillance intervals and the need for endoscopic eradication therapy, as presented in professional society guidelines (Table 1).^4,23,24

Advanced methods for detecting dysplasia

Newer endoscopic surveillance techniques include dye-based chromoendoscopy, narrow-band imaging, confocal laser endomicroscopy, volumetric laser endomicroscopy, and wide-area transepithelial sampling with computer-assisted 3-dimensional analysis. All these techniques are used to increase the detection of dysplasia. Of these, dye-based chromoendoscopy, narrow-band imaging, and confocal laser endomicroscopy meet current criteria of the American Society for Gastrointestinal Endoscopy for preservation and incorporation of valuable endoscopic innovations.²³

MANAGEMENT OF NONDYSPLASTIC BARRETT ESOPHAGUS

A proton pump inhibitor (PPI) is recommended to control reflux symptoms in patients with nondysplastic Barrett esophagus. But it is important to counsel patients on additional ways to protect against esophageal adenocarcinoma, such as:

• Low to moderate alcohol consumption
• Regular physical activity
• Increased dietary intake of fruits, vegetables, folate, fiber, beta-carotene, and vitamin C
• Weight control
• Smoking cessation.²⁵

Surveillance endoscopy with 4-quadrant biopsies at 2-cm intervals is recommended every 3 to 5 years (Table 1).

DOES CHEMOPREVENTION HAVE A ROLE?

Chemoprevention is an exciting area of research in preventing progression to adenocarcinoma in patients with Barrett esophagus. Various drugs such as aspirin, other nonsteroidal anti-inflammatory drugs (NSAIDs), PPIs, metformin, and statins have been studied.

Aspirin

Aspirin has been shown to prevent development of Barrett esophagus in patients with reflux disease,²⁶ but more studies are needed to validate those findings.

PPIs

Gastroesophageal reflux disease is a primary risk factor for esophageal adenocarcinoma, and gastric acid suppression with PPIs reduces cancer risk. PPI therapy is associated with a 71% decrease in the risk of high-grade dysplasia and adenocarcinoma in patients with Barrett esophagus (OR 0.29, 95% CI 0.12–0.79).²⁷ Long-term therapy (> 2 to 3 years) has a higher protective effect (adjusted OR 0.45, 95% CI 0.19–1.06) than short-term therapy (< 2 to 3 years) (adjusted OR 1.09, 95% CI 0.47–2.56).²⁷

NSAIDs

NSAIDs, including aspirin, have been associated with decreased risk of colon, stomach, lung, breast, and esophageal cancer due to their potential to inhibit cyclooxygenase 2 (COX-2) enzymes.

A meta-analysis demonstrated that aspirin and NSAIDs led to a 32% reduction in the risk of adenocarcinoma (OR 0.68, 95% CI 0.56–0.83). The benefit was even greater if the drug was taken daily or more frequently (OR 0.56, 95% CI 0.43–0.73, P < .001) or was taken for 10 or more years (OR 0.63, 95% CI 0.45–0.90, P = .04).²⁸

PPI plus aspirin

The best evidence for the role of PPIs and aspirin in reducing the risk of dysplasia comes from the Aspirin and Esomeprazole Chemoprevention in Barrett’s Metaplasia Trial.²⁹ This randomized, controlled trial compared 4 regimens consisting of esomeprazole (a PPI) in either a high dose (40 mg twice daily) or a low dose (20 mg once daily) plus either aspirin (300 or 320 mg per day) or no aspirin in 2,557 patients with Barrett esophagus. The composite end point was the time to all-cause mortality, adenocarcinoma, or high-grade dysplasia.

At a median follow-up of 8.9 years, the combination of high-dose esomeprazole plus aspirin had the strongest effect compared with low-dose esomeprazole without aspirin (time ratio 1.59, 95% CI 1.14–2.23, P = .0068). The number needed to treat was 34 for esomeprazole and 43 for aspirin.²⁹

Based on these data, we can conclude that aspirin and PPIs can prevent dysplasia and all-cause mortality in Barrett esophagus.

Metformin: No evidence of benefit

Metformin was studied as a protective agent against obesity-associated cancers including esophageal adenocarcinoma, as it reduces insulin levels.

In a randomized controlled trial³⁰ in 74 patients with Barrett esophagus, metformin (starting at 500 mg daily, increasing to 2,000 mg/day by week 4) was compared with placebo. At 12 weeks, the percent change in esophageal levels of the biomarker pS6K1—an intracellular mediator of insulin and insulin-like growth factor activation in Barrett epithelium—did not differ significantly between the 2 groups (1.4% with metformin vs −14.7% with placebo; 1-sided P = .80). This suggested that metformin did not significantly alter proliferation or apoptosis in Barrett epithelium, despite reducing serum insulin levels and insulin resistance. Thus, metformin did not demonstrate a chemoprotective effect in preventing the progression of Barrett esophagus to adenocarcinoma.

Vitamin D affects genes regulating proliferation, apoptosis, and differentiation, and has therefore been studied as a potential antineoplastic agent. Its deficiency has also been associated with increased risk of esophageal adenocarcinoma. However, its efficacy in chemoprevention is unclear.³¹

One study found no association between serum 25-hydroxyvitamin D levels and prevalence of dysplasia in Barrett esophagus (P = .90). An increase in vitamin D levels had no effect on progression to dysplasia or cancer (for every 5-nmol/L increase from baseline, hazard ratio 0.98, P = .62).³²

In another study, supplementation with vitamin D₃ (cholecalciferol 50,000 IU weekly) plus a PPI for 12 weeks significantly improved the serum 25-hydroxyvitamin D levels without significant changes in gene expression from Barrett epithelium.³³ These findings were confirmed in a meta-analysis that showed no consistent association between vitamin D exposure and risk of esophageal neoplasm.³⁴

Thus, there is currently no evidence to support vitamin D for chemoprevention in Barrett esophagus or esophageal adenocarcinoma.

Statins

In addition to lowering cholesterol, statins have antiproliferative, pro-apoptotic, anti-angiogenic, and immunomodulatory effects that prevent cancer, leading to a 41% reduction in the risk of adenocarcinoma in patients with Barrett esophagus in one study (adjusted OR 0.59, 95% CI 0.45–0.78); the number needed to treat with statins to prevent 1 case of adenocarcinoma was 389.³⁵

A meta-analysis also showed that statin use was associated with a lower risk of progression of Barrett esophagus (OR 0.48, 95% CI 0.31–0.73).³⁶

In general, statins appear promising for chemoprevention, but more study is needed.

When is chemoprevention appropriate?

Chemoprevention is not recommended for all patients with Barrett esophagus, given that the condition affects 1% to 2% of the US adult population, and very few patients have progression to esophageal adenocarcinoma. Rather, chemoprevention may be considered in patients with Barrett esophagus and multiple risk factors for adenocarcinoma.

INDEFINITE DYSPLASIA

In Barrett esophagus with indefinite dysplasia, either the epithelial abnormalities are insufficient for a diagnosis of dysplasia, or the nature of the epithelial abnormalities is uncertain due to inflammation or technical difficulties with specimen processing. The risk of high-grade dysplasia or cancer within 1 year of the diagnosis of indefinite dysplasia varies between 1.9% and 15%.³⁷ The recommendation for management is to optimize acid-suppressive therapy for 3 to 6 months and then to repeat esophagogastroduodenoscopy. If indefinite dysplasia is noted again, repeat endoscopy in 12 months is recommended.²

ENDOSCOPIC ERADICATION: AN OVERVIEW

Because dysplasia in Barrett esophagus carries a high risk of progression to cancer, the standard of care is endoscopic mucosal resection of visible lesions, followed by ablation of the flat mucosa, with the aim of achieving complete eradication of intestinal metaplasia.^4,38 The initial endoscopic treatment is followed by outpatient sessions every 8 to 10 weeks until the dysplasia is eradicated. A key part of treatment during this time is maximal acid suppression with a PPI twice daily and a histamine-2 blocker at night. In rare cases, fundoplication is required to control reflux refractory to medical therapy.

After eradication is confirmed, continued surveillance is necessary, as recurrences have been reported at a rate of 4.8% per year for intestinal metaplasia, and 2% per year for dysplasia.³⁹

Current endoscopic resection techniques

Endoscopic resection techniques include mucosal resection, submucosal dissection, radio­frequency ablation, cryotherapy, argon plasma coagulation, and photodynamic therapy (Figure 2).

In mucosal resection, the lesion is either suctioned into a band ligator, after which a band is placed around the lesion, or suctioned into a cap fitted at the end of the endoscope, after which the lesion is removed using a snare.

In submucosal dissection, a liquid is injected into the submucosa to lift the lesion, making it easier to remove. The procedure is technically complex and requires additional training.

In radiofrequency ablation, a special catheter is passed through the endoscope to ablate the affected epithelium by thermal injury. Argon plasma coagulation works in a similar way, but uses ionized argon gas to induce thermal coagulation of metaplastic epithelium.

Cryotherapy produces cellular injury by rapid freezing and thawing of tissue using a cryogen such as liquid nitrogen or nitrous oxide.

In photodynamic therapy, a photosensitizer (porfimer sodium) is administered and taken up preferentially by metaplastic epithelium; it is then activated by transmission of red light using the endoscope, leading to destruction of the metaplastic epithelium.

Of the different techniques, radiofrequency ablation has the most evidence for efficacy and hence is the most commonly used.

All of these procedures are generally well tolerated and have favorable side-effect profiles. After radiofrequency ablation with or without mucosal resection, esophageal strictures are noted in 5.6% of patients, and bleeding and perforation occur rarely (1% and 0.6% of patients, respectively).⁴⁰ Submucosal dissection is associated with a higher rate of complications such as stricture formation (11% of patients) and bleeding or perforation (1.5% of patients).⁴¹

Most patients with low-grade dysplasia (73%) are down-staged to nondysplastic Barrett esophagus or to indefinite for dysplasia after review by expert pathologists.⁴² Patients with confirmed and persistent low-grade dysplasia are at higher risk of progression.⁴³

Once low-grade dysplasia is confirmed by a second gastrointestinal pathologist, the patient should undergo endoscopic ablation. A landmark study by Shaheen et al⁴⁴ demonstrated the benefit of radiofrequency ablation in achieving complete eradication of dysplasia (90.5% vs 22.7% for a sham procedure) and complete eradication of intestinal metaplasia (77.4% vs 2.3% for a sham procedure). In another trial of 136 patients with low-grade dysplasia followed for 3 years, Phoa et al⁴⁵ demonstrated that radiofrequency ablation reduced the rate of progression to high-grade dysplasia by 25% and to adenocarcinoma by 7.4% compared with endoscopic surveillance.

Patients with confirmed low-grade dysplasia who do not undergo eradication therapy should have surveillance endoscopy every 6 to 12 months (Table 1).

HIGH-GRADE DYSPLASIA: RECOMMENDED MANAGEMENT

As with low-grade dysplasia, the diagnosis of high-grade dysplasia needs to be confirmed by a second pathologist with gastrointestinal expertise. In the past, the treatment was esophagectomy, but due to lower morbidity and equivalent efficacy of radiofrequency ablation,⁴⁶ the current treatment of choice is endoscopic mucosal resection of raised lesions, followed by radiofrequency ablation of the entire affected segment.

In the study by Shaheen et al,⁴⁴ 42 patients with high-grade dysplasia were randomized to radiofrequency ablation and 21 to a sham procedure, and 81% of ablation patients achieved complete eradication of dysplasia vs 19% with the sham procedure. Eradication of intestinal metaplasia was achieved in 77% of ablation patients vs 2% of patients with the sham therapy. Results of 3-year follow-up from the same cohort showed complete eradication of dysplasia in 98% and of intestinal metaplasia in 91%.⁴⁷

Endoscopic eradication therapy is recommended for all patients with Barrett esophagus and high-grade dysplasia without a life-limiting comorbidity. Alternatively, surveillance every 3 months is an option if the patient does not wish to undergo eradication therapy. Radiofrequency ablation is more cost-effective than esophagectomy or endoscopic surveillance followed by treatment once patients develop adenocarcinoma.^48,49

EARLY ESOPHAGEAL ADENOCARCINOMA: RECOMMENDED MANAGEMENT

Adenocarcinoma limited to the mucosa and without evidence of nodal involvement can be resected endoscopically. In patients with localized cancer, mucosal resection is done not only for therapeutic purposes but also for staging. Ideal management is multidisciplinary, including a gastroenterologist, thoracic surgeon, oncologist, pathologist, and radiation oncologist.

If lesions have features suggesting submucosal invasion or are greater than 1.5 cm in size, or if it is difficult to separate (ie, lift) the mucosa from the submucosal layer with injection of saline, then submucosal dissection is recommended.⁵⁰ Because of the risk of metachronous lesions, ablation of the remaining Barrett esophagus mucosa is recommended after resection of cancer.

Endoscopic eradication is highly effective and durable for the treatment of intramucosal esophageal adenocarcinoma. In a study of 1,000 patients, 963 patients (96.3%) had achieved a complete response; 12 patients (3.7%) underwent surgery after eradication failed during a follow-up of almost 5 years.⁵¹ Metachronous lesions or recurrence of cancer developed during the follow-up period in 140 patients (14.5%) but were successfully treated endoscopically in 115, resulting in a long-term complete remission rate of 93.8%.

POSTABLATION MANAGEMENT

Because of the risk of recurrence of dysplasia after ablation, long-term PPI therapy and surveillance are recommended.

Surveillance endoscopy involves 4-quadrant biopsies taken every 1 cm from the entire length of segment where Barrett esophagus had been seen before ablation.

The timing of surveillance intervals depends on the preablation grade of dysplasia. For low-grade dysplasia, the recommendation is every 6 months for the first year after ablation and, if there is no recurrence of dysplasia, annually after that.² After treatment of high-grade dysplasia or intramucosal adenocarcinoma, the recommendation is every 3 months for the first year, every 6 months in the second year, and then annually.²

All cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus.¹ But most cases of Barrett esophagus go undiagnosed. And Barrett esophagus is seen in 5% to 15% of patients with gastroesophageal reflux disease.² These facts clearly emphasize the need for screening. Here, we review the rationale and recommendations for screening and surveillance, as well as the range of treatment options.

SCOPE OF THE PROBLEM

The American Cancer Society estimated there were 17,290 new cases of esophageal cancer and 15,850 deaths from it in the United States in 2018.³ Of the 2 main histologic types of esophageal cancer, adenocarcinoma and squamous cell cancer, adenocarcinoma is more common in the United States.

The precursor lesion is Barrett esophagus, defined as an extension of salmon-colored mucosa at least 1 cm into the tubular esophagus proximal to the gastroesophageal junction, with biopsy confirmation of intestinal metaplasia.⁴

The natural course of progression to dysplasia and cancer in Barrett esophagus is unknown but is thought to be stepwise, from no dysplasia to low-grade dysplasia to high-grade dysplasia and cancer, and the cancer risk depends on the degree of dysplasia: the annual risk is 0.33% if there is no dysplasia, 0.54% with low-grade dysplasia, and 7% with high-grade dysplasia.⁴

Although all cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus,¹ more than 90% of patients with newly diagnosed esophageal adenocarcinoma do not have a prior diagnosis of Barrett esophagus.⁵ Therefore, there is a substantial unmet need to expand screening for Barrett esophagus in people at risk.

GASTROESOPHAGEAL REFLUX DISEASE IS A RISK FACTOR FOR CANCER

The rationale behind screening is that detecting Barrett esophagus early and intervening in a timely manner in patients at higher risk of developing adenocarcinoma will decrease mortality.

Chronic gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma (odds ratio [OR] 7.7, 95% confidence interval [CI] 5.3–11.4), and the risk increases when symptoms are long-standing (> 20 years) or severe (OR 43.5, 95% CI 18.3–103.5) or occur daily (OR 5.5, 95% CI 3.2–9.3).⁶

Reflux symptoms are scored as follows:

• Heartburn only, 1 point
• Regurgitation only, 1 point
• Heartburn with regurgitation, 1.5 points
• Nightly symptoms (2 points if yes, 0 if no)
• Symptoms once a week, 0 points; 2 to 6 times a week, 1 point; 7 to 15 times a week, 2 points; more than 15 times a week, 3 points.⁶

A score of 4.5 or higher indicates severe reflux disease. However, it is worth noting that the annual incidence of esophageal adenocarcinoma in patients with long-term gastroesophageal reflux disease is less than 0.001%.⁷

RISK FACTORS FOR BARRETT ESOPHAGUS

Risk factors for Barrett esophagus include:

Male sex. Barrett esophagus is more prevalent in men than in women, at a ratio of 2 to 1; but in individuals under age 50, the ratio is 4 to 1.⁸

Age 50 or older. Barrett esophagus is usually diagnosed in the sixth to seventh decade of life, and the prevalence increases from 2.1% in the third decade to 9.3% in the sixth decade.⁹

White race. It is more prevalent in whites than in blacks (5.0% vs 1.5%, P < .0001).¹⁰

Central obesity. Waist circumference is an independent risk factor: every 5-cm increase carries an OR of 1.14 (95% CI 1.03–1.27, P = .02).¹¹

Cigarette smoking increases the risk of Barrett esophagus (OR 1.42; 95% CI 1.15–1.76).¹²

A family history of Barrett esophagus or esophageal adenocarcinoma is a strong risk factor (OR 12, 95% CI 3.3–44.8). In 1 study, the risk in first- and second-degree relatives of patients with Barrett esophagus was 24%, compared with 5% in a control population (P < .005).¹³

SCREENING GUIDELINES AND DRAWBACKS

The standard screening method is esophagogastroduodenoscopy with sedation, with careful visual inspection and 4-quadrant biopsies every 2 cm using the Seattle protocol, ie, including biopsy of any mucosal irregularities in salmon-colored mucosa above the gastroesophageal junction (Figure 1).⁴

Endoscopic screening is cost-effective, costing $10,440 per quality-adjusted life-year saved, which is well below the accepted threshold of less than $100,000.¹⁴ However, it is still expensive, invasive, and not ideal for screening large populations.

Less-invasive methods under study

Less-invasive, less-expensive methods being tested for mass screening include:

Unsedated transnasal endoscopy. Done with only topical anesthesia, it has high diagnostic accuracy and is quicker and more cost-effective than standard esophagogastroduodenoscopy, with fewer adverse effects. However, the procedure has not yet gained widespread acceptance for regular use by gastroenterologists.¹⁵

A swallowable sponge. Another promising test is cell collection using the Cytosponge Cell Collection Device (Medtronic, Minneapolis, MN). An encapsulated compressed sponge with a string attached is swallowed; in the stomach, the capsule dissolves, and the sponge expands and is then withdrawn using the attached string. The obtained cytology sample from the lower esophagus is then tested for trefoil factor 3, a protein biomarker for Barrett esophagus.¹⁶

A retractable balloon. The EsoCheck Cell Collection Device is a retractable balloon attached to a string. When swallowed, it gathers distal esophageal cells for detecting methylated DNA markers for Barrett esophagus.¹⁷

Esophageal capsule endoscopy uses a camera to visualize the esophagus, but lacks the ability to obtain biopsy samples.

Other screening methods are being tested, although data are limited. Liquid biopsy uses a blood sample to detect microRNAs that are dysregulated in cancer. The “electronic nose” is a device that detects exhaled volatile organic compounds altered in Barrett esophagus. Another test involves taking an oral wash sample to study the oral microbiome for a pattern associated with adenocarcinoma.^18–21

Surveillance in Barrett esophagus aims to detect premalignant changes or early-stage adenocarcinoma to provide longer survival and lower cancer-related mortality. Recent evidence suggests that patients with esophageal adenocarcinoma that is diagnosed in a Barrett esophagus surveillance program have an earlier stage of disease and therefore a survival benefit.²²

Patient education is essential

Before enrolling a patient in a surveillance program, the clinician should explain the risks, benefits, and limitations, the importance of periodic endoscopy, and the possible eventual need for endoscopic therapy or surgery.

The endoscopic procedure

Surveillance involves examination by high-definition white-light endoscopy, with random 4-quadrant biopsies every 2 cm (or every 1 cm in patients with a history of dysplasia) and biopsy of any mucosal irregularity (nodule, ulcer, or other visible lesion). The degree of dysplasia determines the frequency of follow-up surveillance intervals and the need for endoscopic eradication therapy, as presented in professional society guidelines (Table 1).^4,23,24

Advanced methods for detecting dysplasia

Newer endoscopic surveillance techniques include dye-based chromoendoscopy, narrow-band imaging, confocal laser endomicroscopy, volumetric laser endomicroscopy, and wide-area transepithelial sampling with computer-assisted 3-dimensional analysis. All these techniques are used to increase the detection of dysplasia. Of these, dye-based chromoendoscopy, narrow-band imaging, and confocal laser endomicroscopy meet current criteria of the American Society for Gastrointestinal Endoscopy for preservation and incorporation of valuable endoscopic innovations.²³

MANAGEMENT OF NONDYSPLASTIC BARRETT ESOPHAGUS

A proton pump inhibitor (PPI) is recommended to control reflux symptoms in patients with nondysplastic Barrett esophagus. But it is important to counsel patients on additional ways to protect against esophageal adenocarcinoma, such as:

• Low to moderate alcohol consumption
• Regular physical activity
• Increased dietary intake of fruits, vegetables, folate, fiber, beta-carotene, and vitamin C
• Weight control
• Smoking cessation.²⁵

Surveillance endoscopy with 4-quadrant biopsies at 2-cm intervals is recommended every 3 to 5 years (Table 1).

DOES CHEMOPREVENTION HAVE A ROLE?

Chemoprevention is an exciting area of research in preventing progression to adenocarcinoma in patients with Barrett esophagus. Various drugs such as aspirin, other nonsteroidal anti-inflammatory drugs (NSAIDs), PPIs, metformin, and statins have been studied.

Aspirin

Aspirin has been shown to prevent development of Barrett esophagus in patients with reflux disease,²⁶ but more studies are needed to validate those findings.

PPIs

Gastroesophageal reflux disease is a primary risk factor for esophageal adenocarcinoma, and gastric acid suppression with PPIs reduces cancer risk. PPI therapy is associated with a 71% decrease in the risk of high-grade dysplasia and adenocarcinoma in patients with Barrett esophagus (OR 0.29, 95% CI 0.12–0.79).²⁷ Long-term therapy (> 2 to 3 years) has a higher protective effect (adjusted OR 0.45, 95% CI 0.19–1.06) than short-term therapy (< 2 to 3 years) (adjusted OR 1.09, 95% CI 0.47–2.56).²⁷

NSAIDs

NSAIDs, including aspirin, have been associated with decreased risk of colon, stomach, lung, breast, and esophageal cancer due to their potential to inhibit cyclooxygenase 2 (COX-2) enzymes.

A meta-analysis demonstrated that aspirin and NSAIDs led to a 32% reduction in the risk of adenocarcinoma (OR 0.68, 95% CI 0.56–0.83). The benefit was even greater if the drug was taken daily or more frequently (OR 0.56, 95% CI 0.43–0.73, P < .001) or was taken for 10 or more years (OR 0.63, 95% CI 0.45–0.90, P = .04).²⁸

PPI plus aspirin

The best evidence for the role of PPIs and aspirin in reducing the risk of dysplasia comes from the Aspirin and Esomeprazole Chemoprevention in Barrett’s Metaplasia Trial.²⁹ This randomized, controlled trial compared 4 regimens consisting of esomeprazole (a PPI) in either a high dose (40 mg twice daily) or a low dose (20 mg once daily) plus either aspirin (300 or 320 mg per day) or no aspirin in 2,557 patients with Barrett esophagus. The composite end point was the time to all-cause mortality, adenocarcinoma, or high-grade dysplasia.

At a median follow-up of 8.9 years, the combination of high-dose esomeprazole plus aspirin had the strongest effect compared with low-dose esomeprazole without aspirin (time ratio 1.59, 95% CI 1.14–2.23, P = .0068). The number needed to treat was 34 for esomeprazole and 43 for aspirin.²⁹

Based on these data, we can conclude that aspirin and PPIs can prevent dysplasia and all-cause mortality in Barrett esophagus.

Metformin: No evidence of benefit

Metformin was studied as a protective agent against obesity-associated cancers including esophageal adenocarcinoma, as it reduces insulin levels.

In a randomized controlled trial³⁰ in 74 patients with Barrett esophagus, metformin (starting at 500 mg daily, increasing to 2,000 mg/day by week 4) was compared with placebo. At 12 weeks, the percent change in esophageal levels of the biomarker pS6K1—an intracellular mediator of insulin and insulin-like growth factor activation in Barrett epithelium—did not differ significantly between the 2 groups (1.4% with metformin vs −14.7% with placebo; 1-sided P = .80). This suggested that metformin did not significantly alter proliferation or apoptosis in Barrett epithelium, despite reducing serum insulin levels and insulin resistance. Thus, metformin did not demonstrate a chemoprotective effect in preventing the progression of Barrett esophagus to adenocarcinoma.

Vitamin D affects genes regulating proliferation, apoptosis, and differentiation, and has therefore been studied as a potential antineoplastic agent. Its deficiency has also been associated with increased risk of esophageal adenocarcinoma. However, its efficacy in chemoprevention is unclear.³¹

One study found no association between serum 25-hydroxyvitamin D levels and prevalence of dysplasia in Barrett esophagus (P = .90). An increase in vitamin D levels had no effect on progression to dysplasia or cancer (for every 5-nmol/L increase from baseline, hazard ratio 0.98, P = .62).³²

In another study, supplementation with vitamin D₃ (cholecalciferol 50,000 IU weekly) plus a PPI for 12 weeks significantly improved the serum 25-hydroxyvitamin D levels without significant changes in gene expression from Barrett epithelium.³³ These findings were confirmed in a meta-analysis that showed no consistent association between vitamin D exposure and risk of esophageal neoplasm.³⁴

Thus, there is currently no evidence to support vitamin D for chemoprevention in Barrett esophagus or esophageal adenocarcinoma.

Statins

In addition to lowering cholesterol, statins have antiproliferative, pro-apoptotic, anti-angiogenic, and immunomodulatory effects that prevent cancer, leading to a 41% reduction in the risk of adenocarcinoma in patients with Barrett esophagus in one study (adjusted OR 0.59, 95% CI 0.45–0.78); the number needed to treat with statins to prevent 1 case of adenocarcinoma was 389.³⁵

A meta-analysis also showed that statin use was associated with a lower risk of progression of Barrett esophagus (OR 0.48, 95% CI 0.31–0.73).³⁶

In general, statins appear promising for chemoprevention, but more study is needed.

When is chemoprevention appropriate?

Chemoprevention is not recommended for all patients with Barrett esophagus, given that the condition affects 1% to 2% of the US adult population, and very few patients have progression to esophageal adenocarcinoma. Rather, chemoprevention may be considered in patients with Barrett esophagus and multiple risk factors for adenocarcinoma.

INDEFINITE DYSPLASIA

In Barrett esophagus with indefinite dysplasia, either the epithelial abnormalities are insufficient for a diagnosis of dysplasia, or the nature of the epithelial abnormalities is uncertain due to inflammation or technical difficulties with specimen processing. The risk of high-grade dysplasia or cancer within 1 year of the diagnosis of indefinite dysplasia varies between 1.9% and 15%.³⁷ The recommendation for management is to optimize acid-suppressive therapy for 3 to 6 months and then to repeat esophagogastroduodenoscopy. If indefinite dysplasia is noted again, repeat endoscopy in 12 months is recommended.²

ENDOSCOPIC ERADICATION: AN OVERVIEW

Because dysplasia in Barrett esophagus carries a high risk of progression to cancer, the standard of care is endoscopic mucosal resection of visible lesions, followed by ablation of the flat mucosa, with the aim of achieving complete eradication of intestinal metaplasia.^4,38 The initial endoscopic treatment is followed by outpatient sessions every 8 to 10 weeks until the dysplasia is eradicated. A key part of treatment during this time is maximal acid suppression with a PPI twice daily and a histamine-2 blocker at night. In rare cases, fundoplication is required to control reflux refractory to medical therapy.

After eradication is confirmed, continued surveillance is necessary, as recurrences have been reported at a rate of 4.8% per year for intestinal metaplasia, and 2% per year for dysplasia.³⁹

Current endoscopic resection techniques

Endoscopic resection techniques include mucosal resection, submucosal dissection, radio­frequency ablation, cryotherapy, argon plasma coagulation, and photodynamic therapy (Figure 2).

In mucosal resection, the lesion is either suctioned into a band ligator, after which a band is placed around the lesion, or suctioned into a cap fitted at the end of the endoscope, after which the lesion is removed using a snare.

In submucosal dissection, a liquid is injected into the submucosa to lift the lesion, making it easier to remove. The procedure is technically complex and requires additional training.

In radiofrequency ablation, a special catheter is passed through the endoscope to ablate the affected epithelium by thermal injury. Argon plasma coagulation works in a similar way, but uses ionized argon gas to induce thermal coagulation of metaplastic epithelium.

Cryotherapy produces cellular injury by rapid freezing and thawing of tissue using a cryogen such as liquid nitrogen or nitrous oxide.

In photodynamic therapy, a photosensitizer (porfimer sodium) is administered and taken up preferentially by metaplastic epithelium; it is then activated by transmission of red light using the endoscope, leading to destruction of the metaplastic epithelium.

Of the different techniques, radiofrequency ablation has the most evidence for efficacy and hence is the most commonly used.

All of these procedures are generally well tolerated and have favorable side-effect profiles. After radiofrequency ablation with or without mucosal resection, esophageal strictures are noted in 5.6% of patients, and bleeding and perforation occur rarely (1% and 0.6% of patients, respectively).⁴⁰ Submucosal dissection is associated with a higher rate of complications such as stricture formation (11% of patients) and bleeding or perforation (1.5% of patients).⁴¹

Most patients with low-grade dysplasia (73%) are down-staged to nondysplastic Barrett esophagus or to indefinite for dysplasia after review by expert pathologists.⁴² Patients with confirmed and persistent low-grade dysplasia are at higher risk of progression.⁴³

Once low-grade dysplasia is confirmed by a second gastrointestinal pathologist, the patient should undergo endoscopic ablation. A landmark study by Shaheen et al⁴⁴ demonstrated the benefit of radiofrequency ablation in achieving complete eradication of dysplasia (90.5% vs 22.7% for a sham procedure) and complete eradication of intestinal metaplasia (77.4% vs 2.3% for a sham procedure). In another trial of 136 patients with low-grade dysplasia followed for 3 years, Phoa et al⁴⁵ demonstrated that radiofrequency ablation reduced the rate of progression to high-grade dysplasia by 25% and to adenocarcinoma by 7.4% compared with endoscopic surveillance.

Patients with confirmed low-grade dysplasia who do not undergo eradication therapy should have surveillance endoscopy every 6 to 12 months (Table 1).

HIGH-GRADE DYSPLASIA: RECOMMENDED MANAGEMENT

As with low-grade dysplasia, the diagnosis of high-grade dysplasia needs to be confirmed by a second pathologist with gastrointestinal expertise. In the past, the treatment was esophagectomy, but due to lower morbidity and equivalent efficacy of radiofrequency ablation,⁴⁶ the current treatment of choice is endoscopic mucosal resection of raised lesions, followed by radiofrequency ablation of the entire affected segment.

In the study by Shaheen et al,⁴⁴ 42 patients with high-grade dysplasia were randomized to radiofrequency ablation and 21 to a sham procedure, and 81% of ablation patients achieved complete eradication of dysplasia vs 19% with the sham procedure. Eradication of intestinal metaplasia was achieved in 77% of ablation patients vs 2% of patients with the sham therapy. Results of 3-year follow-up from the same cohort showed complete eradication of dysplasia in 98% and of intestinal metaplasia in 91%.⁴⁷

Endoscopic eradication therapy is recommended for all patients with Barrett esophagus and high-grade dysplasia without a life-limiting comorbidity. Alternatively, surveillance every 3 months is an option if the patient does not wish to undergo eradication therapy. Radiofrequency ablation is more cost-effective than esophagectomy or endoscopic surveillance followed by treatment once patients develop adenocarcinoma.^48,49

EARLY ESOPHAGEAL ADENOCARCINOMA: RECOMMENDED MANAGEMENT

Adenocarcinoma limited to the mucosa and without evidence of nodal involvement can be resected endoscopically. In patients with localized cancer, mucosal resection is done not only for therapeutic purposes but also for staging. Ideal management is multidisciplinary, including a gastroenterologist, thoracic surgeon, oncologist, pathologist, and radiation oncologist.

If lesions have features suggesting submucosal invasion or are greater than 1.5 cm in size, or if it is difficult to separate (ie, lift) the mucosa from the submucosal layer with injection of saline, then submucosal dissection is recommended.⁵⁰ Because of the risk of metachronous lesions, ablation of the remaining Barrett esophagus mucosa is recommended after resection of cancer.

Endoscopic eradication is highly effective and durable for the treatment of intramucosal esophageal adenocarcinoma. In a study of 1,000 patients, 963 patients (96.3%) had achieved a complete response; 12 patients (3.7%) underwent surgery after eradication failed during a follow-up of almost 5 years.⁵¹ Metachronous lesions or recurrence of cancer developed during the follow-up period in 140 patients (14.5%) but were successfully treated endoscopically in 115, resulting in a long-term complete remission rate of 93.8%.

POSTABLATION MANAGEMENT

Because of the risk of recurrence of dysplasia after ablation, long-term PPI therapy and surveillance are recommended.

Surveillance endoscopy involves 4-quadrant biopsies taken every 1 cm from the entire length of segment where Barrett esophagus had been seen before ablation.

The timing of surveillance intervals depends on the preablation grade of dysplasia. For low-grade dysplasia, the recommendation is every 6 months for the first year after ablation and, if there is no recurrence of dysplasia, annually after that.² After treatment of high-grade dysplasia or intramucosal adenocarcinoma, the recommendation is every 3 months for the first year, every 6 months in the second year, and then annually.²

All cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus.¹ But most cases of Barrett esophagus go undiagnosed. And Barrett esophagus is seen in 5% to 15% of patients with gastroesophageal reflux disease.² These facts clearly emphasize the need for screening. Here, we review the rationale and recommendations for screening and surveillance, as well as the range of treatment options.

SCOPE OF THE PROBLEM

The American Cancer Society estimated there were 17,290 new cases of esophageal cancer and 15,850 deaths from it in the United States in 2018.³ Of the 2 main histologic types of esophageal cancer, adenocarcinoma and squamous cell cancer, adenocarcinoma is more common in the United States.

The precursor lesion is Barrett esophagus, defined as an extension of salmon-colored mucosa at least 1 cm into the tubular esophagus proximal to the gastroesophageal junction, with biopsy confirmation of intestinal metaplasia.⁴

The natural course of progression to dysplasia and cancer in Barrett esophagus is unknown but is thought to be stepwise, from no dysplasia to low-grade dysplasia to high-grade dysplasia and cancer, and the cancer risk depends on the degree of dysplasia: the annual risk is 0.33% if there is no dysplasia, 0.54% with low-grade dysplasia, and 7% with high-grade dysplasia.⁴

Although all cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus,¹ more than 90% of patients with newly diagnosed esophageal adenocarcinoma do not have a prior diagnosis of Barrett esophagus.⁵ Therefore, there is a substantial unmet need to expand screening for Barrett esophagus in people at risk.

GASTROESOPHAGEAL REFLUX DISEASE IS A RISK FACTOR FOR CANCER

The rationale behind screening is that detecting Barrett esophagus early and intervening in a timely manner in patients at higher risk of developing adenocarcinoma will decrease mortality.

Chronic gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma (odds ratio [OR] 7.7, 95% confidence interval [CI] 5.3–11.4), and the risk increases when symptoms are long-standing (> 20 years) or severe (OR 43.5, 95% CI 18.3–103.5) or occur daily (OR 5.5, 95% CI 3.2–9.3).⁶

Reflux symptoms are scored as follows:

• Heartburn only, 1 point
• Regurgitation only, 1 point
• Heartburn with regurgitation, 1.5 points
• Nightly symptoms (2 points if yes, 0 if no)
• Symptoms once a week, 0 points; 2 to 6 times a week, 1 point; 7 to 15 times a week, 2 points; more than 15 times a week, 3 points.⁶

A score of 4.5 or higher indicates severe reflux disease. However, it is worth noting that the annual incidence of esophageal adenocarcinoma in patients with long-term gastroesophageal reflux disease is less than 0.001%.⁷

RISK FACTORS FOR BARRETT ESOPHAGUS

Risk factors for Barrett esophagus include:

Male sex. Barrett esophagus is more prevalent in men than in women, at a ratio of 2 to 1; but in individuals under age 50, the ratio is 4 to 1.⁸

Age 50 or older. Barrett esophagus is usually diagnosed in the sixth to seventh decade of life, and the prevalence increases from 2.1% in the third decade to 9.3% in the sixth decade.⁹

White race. It is more prevalent in whites than in blacks (5.0% vs 1.5%, P < .0001).¹⁰

Central obesity. Waist circumference is an independent risk factor: every 5-cm increase carries an OR of 1.14 (95% CI 1.03–1.27, P = .02).¹¹

Cigarette smoking increases the risk of Barrett esophagus (OR 1.42; 95% CI 1.15–1.76).¹²

A family history of Barrett esophagus or esophageal adenocarcinoma is a strong risk factor (OR 12, 95% CI 3.3–44.8). In 1 study, the risk in first- and second-degree relatives of patients with Barrett esophagus was 24%, compared with 5% in a control population (P < .005).¹³

SCREENING GUIDELINES AND DRAWBACKS

The standard screening method is esophagogastroduodenoscopy with sedation, with careful visual inspection and 4-quadrant biopsies every 2 cm using the Seattle protocol, ie, including biopsy of any mucosal irregularities in salmon-colored mucosa above the gastroesophageal junction (Figure 1).⁴

Endoscopic screening is cost-effective, costing $10,440 per quality-adjusted life-year saved, which is well below the accepted threshold of less than $100,000.¹⁴ However, it is still expensive, invasive, and not ideal for screening large populations.

Less-invasive methods under study

Less-invasive, less-expensive methods being tested for mass screening include:

Unsedated transnasal endoscopy. Done with only topical anesthesia, it has high diagnostic accuracy and is quicker and more cost-effective than standard esophagogastroduodenoscopy, with fewer adverse effects. However, the procedure has not yet gained widespread acceptance for regular use by gastroenterologists.¹⁵

A swallowable sponge. Another promising test is cell collection using the Cytosponge Cell Collection Device (Medtronic, Minneapolis, MN). An encapsulated compressed sponge with a string attached is swallowed; in the stomach, the capsule dissolves, and the sponge expands and is then withdrawn using the attached string. The obtained cytology sample from the lower esophagus is then tested for trefoil factor 3, a protein biomarker for Barrett esophagus.¹⁶

A retractable balloon. The EsoCheck Cell Collection Device is a retractable balloon attached to a string. When swallowed, it gathers distal esophageal cells for detecting methylated DNA markers for Barrett esophagus.¹⁷

Esophageal capsule endoscopy uses a camera to visualize the esophagus, but lacks the ability to obtain biopsy samples.

Other screening methods are being tested, although data are limited. Liquid biopsy uses a blood sample to detect microRNAs that are dysregulated in cancer. The “electronic nose” is a device that detects exhaled volatile organic compounds altered in Barrett esophagus. Another test involves taking an oral wash sample to study the oral microbiome for a pattern associated with adenocarcinoma.^18–21

Surveillance in Barrett esophagus aims to detect premalignant changes or early-stage adenocarcinoma to provide longer survival and lower cancer-related mortality. Recent evidence suggests that patients with esophageal adenocarcinoma that is diagnosed in a Barrett esophagus surveillance program have an earlier stage of disease and therefore a survival benefit.²²

Patient education is essential

Before enrolling a patient in a surveillance program, the clinician should explain the risks, benefits, and limitations, the importance of periodic endoscopy, and the possible eventual need for endoscopic therapy or surgery.

The endoscopic procedure

Surveillance involves examination by high-definition white-light endoscopy, with random 4-quadrant biopsies every 2 cm (or every 1 cm in patients with a history of dysplasia) and biopsy of any mucosal irregularity (nodule, ulcer, or other visible lesion). The degree of dysplasia determines the frequency of follow-up surveillance intervals and the need for endoscopic eradication therapy, as presented in professional society guidelines (Table 1).^4,23,24

Advanced methods for detecting dysplasia

Newer endoscopic surveillance techniques include dye-based chromoendoscopy, narrow-band imaging, confocal laser endomicroscopy, volumetric laser endomicroscopy, and wide-area transepithelial sampling with computer-assisted 3-dimensional analysis. All these techniques are used to increase the detection of dysplasia. Of these, dye-based chromoendoscopy, narrow-band imaging, and confocal laser endomicroscopy meet current criteria of the American Society for Gastrointestinal Endoscopy for preservation and incorporation of valuable endoscopic innovations.²³

MANAGEMENT OF NONDYSPLASTIC BARRETT ESOPHAGUS

A proton pump inhibitor (PPI) is recommended to control reflux symptoms in patients with nondysplastic Barrett esophagus. But it is important to counsel patients on additional ways to protect against esophageal adenocarcinoma, such as:

• Low to moderate alcohol consumption
• Regular physical activity
• Increased dietary intake of fruits, vegetables, folate, fiber, beta-carotene, and vitamin C
• Weight control
• Smoking cessation.²⁵

Surveillance endoscopy with 4-quadrant biopsies at 2-cm intervals is recommended every 3 to 5 years (Table 1).

DOES CHEMOPREVENTION HAVE A ROLE?

Chemoprevention is an exciting area of research in preventing progression to adenocarcinoma in patients with Barrett esophagus. Various drugs such as aspirin, other nonsteroidal anti-inflammatory drugs (NSAIDs), PPIs, metformin, and statins have been studied.

Aspirin

Aspirin has been shown to prevent development of Barrett esophagus in patients with reflux disease,²⁶ but more studies are needed to validate those findings.

PPIs

Gastroesophageal reflux disease is a primary risk factor for esophageal adenocarcinoma, and gastric acid suppression with PPIs reduces cancer risk. PPI therapy is associated with a 71% decrease in the risk of high-grade dysplasia and adenocarcinoma in patients with Barrett esophagus (OR 0.29, 95% CI 0.12–0.79).²⁷ Long-term therapy (> 2 to 3 years) has a higher protective effect (adjusted OR 0.45, 95% CI 0.19–1.06) than short-term therapy (< 2 to 3 years) (adjusted OR 1.09, 95% CI 0.47–2.56).²⁷

NSAIDs

NSAIDs, including aspirin, have been associated with decreased risk of colon, stomach, lung, breast, and esophageal cancer due to their potential to inhibit cyclooxygenase 2 (COX-2) enzymes.

A meta-analysis demonstrated that aspirin and NSAIDs led to a 32% reduction in the risk of adenocarcinoma (OR 0.68, 95% CI 0.56–0.83). The benefit was even greater if the drug was taken daily or more frequently (OR 0.56, 95% CI 0.43–0.73, P < .001) or was taken for 10 or more years (OR 0.63, 95% CI 0.45–0.90, P = .04).²⁸

PPI plus aspirin

The best evidence for the role of PPIs and aspirin in reducing the risk of dysplasia comes from the Aspirin and Esomeprazole Chemoprevention in Barrett’s Metaplasia Trial.²⁹ This randomized, controlled trial compared 4 regimens consisting of esomeprazole (a PPI) in either a high dose (40 mg twice daily) or a low dose (20 mg once daily) plus either aspirin (300 or 320 mg per day) or no aspirin in 2,557 patients with Barrett esophagus. The composite end point was the time to all-cause mortality, adenocarcinoma, or high-grade dysplasia.

At a median follow-up of 8.9 years, the combination of high-dose esomeprazole plus aspirin had the strongest effect compared with low-dose esomeprazole without aspirin (time ratio 1.59, 95% CI 1.14–2.23, P = .0068). The number needed to treat was 34 for esomeprazole and 43 for aspirin.²⁹

Based on these data, we can conclude that aspirin and PPIs can prevent dysplasia and all-cause mortality in Barrett esophagus.

Metformin: No evidence of benefit

Metformin was studied as a protective agent against obesity-associated cancers including esophageal adenocarcinoma, as it reduces insulin levels.

In a randomized controlled trial³⁰ in 74 patients with Barrett esophagus, metformin (starting at 500 mg daily, increasing to 2,000 mg/day by week 4) was compared with placebo. At 12 weeks, the percent change in esophageal levels of the biomarker pS6K1—an intracellular mediator of insulin and insulin-like growth factor activation in Barrett epithelium—did not differ significantly between the 2 groups (1.4% with metformin vs −14.7% with placebo; 1-sided P = .80). This suggested that metformin did not significantly alter proliferation or apoptosis in Barrett epithelium, despite reducing serum insulin levels and insulin resistance. Thus, metformin did not demonstrate a chemoprotective effect in preventing the progression of Barrett esophagus to adenocarcinoma.

Vitamin D affects genes regulating proliferation, apoptosis, and differentiation, and has therefore been studied as a potential antineoplastic agent. Its deficiency has also been associated with increased risk of esophageal adenocarcinoma. However, its efficacy in chemoprevention is unclear.³¹

One study found no association between serum 25-hydroxyvitamin D levels and prevalence of dysplasia in Barrett esophagus (P = .90). An increase in vitamin D levels had no effect on progression to dysplasia or cancer (for every 5-nmol/L increase from baseline, hazard ratio 0.98, P = .62).³²

In another study, supplementation with vitamin D₃ (cholecalciferol 50,000 IU weekly) plus a PPI for 12 weeks significantly improved the serum 25-hydroxyvitamin D levels without significant changes in gene expression from Barrett epithelium.³³ These findings were confirmed in a meta-analysis that showed no consistent association between vitamin D exposure and risk of esophageal neoplasm.³⁴

Thus, there is currently no evidence to support vitamin D for chemoprevention in Barrett esophagus or esophageal adenocarcinoma.

Statins

In addition to lowering cholesterol, statins have antiproliferative, pro-apoptotic, anti-angiogenic, and immunomodulatory effects that prevent cancer, leading to a 41% reduction in the risk of adenocarcinoma in patients with Barrett esophagus in one study (adjusted OR 0.59, 95% CI 0.45–0.78); the number needed to treat with statins to prevent 1 case of adenocarcinoma was 389.³⁵

A meta-analysis also showed that statin use was associated with a lower risk of progression of Barrett esophagus (OR 0.48, 95% CI 0.31–0.73).³⁶

In general, statins appear promising for chemoprevention, but more study is needed.

When is chemoprevention appropriate?

Chemoprevention is not recommended for all patients with Barrett esophagus, given that the condition affects 1% to 2% of the US adult population, and very few patients have progression to esophageal adenocarcinoma. Rather, chemoprevention may be considered in patients with Barrett esophagus and multiple risk factors for adenocarcinoma.

INDEFINITE DYSPLASIA

In Barrett esophagus with indefinite dysplasia, either the epithelial abnormalities are insufficient for a diagnosis of dysplasia, or the nature of the epithelial abnormalities is uncertain due to inflammation or technical difficulties with specimen processing. The risk of high-grade dysplasia or cancer within 1 year of the diagnosis of indefinite dysplasia varies between 1.9% and 15%.³⁷ The recommendation for management is to optimize acid-suppressive therapy for 3 to 6 months and then to repeat esophagogastroduodenoscopy. If indefinite dysplasia is noted again, repeat endoscopy in 12 months is recommended.²

ENDOSCOPIC ERADICATION: AN OVERVIEW

Because dysplasia in Barrett esophagus carries a high risk of progression to cancer, the standard of care is endoscopic mucosal resection of visible lesions, followed by ablation of the flat mucosa, with the aim of achieving complete eradication of intestinal metaplasia.^4,38 The initial endoscopic treatment is followed by outpatient sessions every 8 to 10 weeks until the dysplasia is eradicated. A key part of treatment during this time is maximal acid suppression with a PPI twice daily and a histamine-2 blocker at night. In rare cases, fundoplication is required to control reflux refractory to medical therapy.

After eradication is confirmed, continued surveillance is necessary, as recurrences have been reported at a rate of 4.8% per year for intestinal metaplasia, and 2% per year for dysplasia.³⁹

Current endoscopic resection techniques

Endoscopic resection techniques include mucosal resection, submucosal dissection, radio­frequency ablation, cryotherapy, argon plasma coagulation, and photodynamic therapy (Figure 2).

In mucosal resection, the lesion is either suctioned into a band ligator, after which a band is placed around the lesion, or suctioned into a cap fitted at the end of the endoscope, after which the lesion is removed using a snare.

In submucosal dissection, a liquid is injected into the submucosa to lift the lesion, making it easier to remove. The procedure is technically complex and requires additional training.

In radiofrequency ablation, a special catheter is passed through the endoscope to ablate the affected epithelium by thermal injury. Argon plasma coagulation works in a similar way, but uses ionized argon gas to induce thermal coagulation of metaplastic epithelium.

Cryotherapy produces cellular injury by rapid freezing and thawing of tissue using a cryogen such as liquid nitrogen or nitrous oxide.

In photodynamic therapy, a photosensitizer (porfimer sodium) is administered and taken up preferentially by metaplastic epithelium; it is then activated by transmission of red light using the endoscope, leading to destruction of the metaplastic epithelium.

Of the different techniques, radiofrequency ablation has the most evidence for efficacy and hence is the most commonly used.

All of these procedures are generally well tolerated and have favorable side-effect profiles. After radiofrequency ablation with or without mucosal resection, esophageal strictures are noted in 5.6% of patients, and bleeding and perforation occur rarely (1% and 0.6% of patients, respectively).⁴⁰ Submucosal dissection is associated with a higher rate of complications such as stricture formation (11% of patients) and bleeding or perforation (1.5% of patients).⁴¹

Most patients with low-grade dysplasia (73%) are down-staged to nondysplastic Barrett esophagus or to indefinite for dysplasia after review by expert pathologists.⁴² Patients with confirmed and persistent low-grade dysplasia are at higher risk of progression.⁴³

Once low-grade dysplasia is confirmed by a second gastrointestinal pathologist, the patient should undergo endoscopic ablation. A landmark study by Shaheen et al⁴⁴ demonstrated the benefit of radiofrequency ablation in achieving complete eradication of dysplasia (90.5% vs 22.7% for a sham procedure) and complete eradication of intestinal metaplasia (77.4% vs 2.3% for a sham procedure). In another trial of 136 patients with low-grade dysplasia followed for 3 years, Phoa et al⁴⁵ demonstrated that radiofrequency ablation reduced the rate of progression to high-grade dysplasia by 25% and to adenocarcinoma by 7.4% compared with endoscopic surveillance.

Patients with confirmed low-grade dysplasia who do not undergo eradication therapy should have surveillance endoscopy every 6 to 12 months (Table 1).

HIGH-GRADE DYSPLASIA: RECOMMENDED MANAGEMENT

As with low-grade dysplasia, the diagnosis of high-grade dysplasia needs to be confirmed by a second pathologist with gastrointestinal expertise. In the past, the treatment was esophagectomy, but due to lower morbidity and equivalent efficacy of radiofrequency ablation,⁴⁶ the current treatment of choice is endoscopic mucosal resection of raised lesions, followed by radiofrequency ablation of the entire affected segment.

In the study by Shaheen et al,⁴⁴ 42 patients with high-grade dysplasia were randomized to radiofrequency ablation and 21 to a sham procedure, and 81% of ablation patients achieved complete eradication of dysplasia vs 19% with the sham procedure. Eradication of intestinal metaplasia was achieved in 77% of ablation patients vs 2% of patients with the sham therapy. Results of 3-year follow-up from the same cohort showed complete eradication of dysplasia in 98% and of intestinal metaplasia in 91%.⁴⁷

Endoscopic eradication therapy is recommended for all patients with Barrett esophagus and high-grade dysplasia without a life-limiting comorbidity. Alternatively, surveillance every 3 months is an option if the patient does not wish to undergo eradication therapy. Radiofrequency ablation is more cost-effective than esophagectomy or endoscopic surveillance followed by treatment once patients develop adenocarcinoma.^48,49

EARLY ESOPHAGEAL ADENOCARCINOMA: RECOMMENDED MANAGEMENT

Adenocarcinoma limited to the mucosa and without evidence of nodal involvement can be resected endoscopically. In patients with localized cancer, mucosal resection is done not only for therapeutic purposes but also for staging. Ideal management is multidisciplinary, including a gastroenterologist, thoracic surgeon, oncologist, pathologist, and radiation oncologist.

If lesions have features suggesting submucosal invasion or are greater than 1.5 cm in size, or if it is difficult to separate (ie, lift) the mucosa from the submucosal layer with injection of saline, then submucosal dissection is recommended.⁵⁰ Because of the risk of metachronous lesions, ablation of the remaining Barrett esophagus mucosa is recommended after resection of cancer.

Endoscopic eradication is highly effective and durable for the treatment of intramucosal esophageal adenocarcinoma. In a study of 1,000 patients, 963 patients (96.3%) had achieved a complete response; 12 patients (3.7%) underwent surgery after eradication failed during a follow-up of almost 5 years.⁵¹ Metachronous lesions or recurrence of cancer developed during the follow-up period in 140 patients (14.5%) but were successfully treated endoscopically in 115, resulting in a long-term complete remission rate of 93.8%.

POSTABLATION MANAGEMENT

Because of the risk of recurrence of dysplasia after ablation, long-term PPI therapy and surveillance are recommended.

Surveillance endoscopy involves 4-quadrant biopsies taken every 1 cm from the entire length of segment where Barrett esophagus had been seen before ablation.

The timing of surveillance intervals depends on the preablation grade of dysplasia. For low-grade dysplasia, the recommendation is every 6 months for the first year after ablation and, if there is no recurrence of dysplasia, annually after that.² After treatment of high-grade dysplasia or intramucosal adenocarcinoma, the recommendation is every 3 months for the first year, every 6 months in the second year, and then annually.²

Staging of liver fibrosis, important for determining prognosis in patients with chronic liver disease and for the need to start screening for complications of cirrhosis, was traditionally done only by liver biopsy. While biopsy is still the gold standard method to stage fibrosis, noninvasive methods have been developed that can also assess disease severity.

This article briefly reviews the epidemiology and physiology of chronic liver disease and the traditional role of liver biopsy. Pros and cons of alternative fibrosis assessment methods are discussed, with a focus on their utility for patients with nonalcoholic fatty liver disease (NAFLD) and hepatitis C virus (HCV) infection.

CHRONIC LIVER DISEASE: A HUGE HEALTH BURDEN

Chronic liver disease is associated with enormous health and financial costs in the United States. Its prevalence is about 15%,¹ and it is the 12th leading cause of death.² Hospital costs are estimated at about $4 billion annually.³

The most common causes of chronic liver disease are NAFLD (which may be present in up to one-third of the US population and is increasing with the epidemic of obesity), its aggressive variant, nonalcoholic steatohepatitis (NASH) (present in about 3% of the population), and HCV infection (1%).^4,5

Since direct-acting antiviral agents were introduced, HCV infection dropped from being the leading cause of liver transplant to third place.⁶ But at the same time, the number of patients on the transplant waiting list who have NASH has risen faster than for any other cause of chronic liver disease.⁷

FIBROSIS: A KEY INDICATOR OF DISEASE SEVERITY

In HCV infection, advanced fibrosis is defined as either stage 4 to 6 using the Ishak system¹⁰ or stage 3 to 4 using the Meta-analysis of Histological Data in Viral Hepatitis (METAVIR) system.¹¹

In NAFLD, advanced fibrosis is defined as stage 3 to 4 using the NASH Clinical Research Network system.¹²

Staging fibrosis is also important so that patients with cirrhosis can be identified early to begin screening for hepatocellular carcinoma and esophageal varices to reduce the risks of illness and death. In addition, insurance companies often require documentation of fibrosis stage before treating HCV with the new direct-acting antiviral agents.

LIVER BIOPSY IS STILL THE GOLD STANDARD

Although invasive, liver biopsy remains the gold standard for determining fibrosis stage. Liver biopsies were performed “blindly” (without imaging) until the 1990s, but imaging-guided biopsy using ultrasonography was then developed, which entailed less pain and lower complication and hospitalization rates. Slightly more hepatic tissue is obtained with guided liver biopsy, but the difference was deemed clinically insignificant.¹³ Concern initially arose about the added cost involved with imaging, but imaging-guided biopsy was actually found to be more cost-effective.¹⁴

In the 2000s, transjugular liver biopsy via the right internal jugular vein became available. This method was originally used primarily in patients with ascites or significant coagulopathy. At first, there were concerns about the adequacy of specimens obtained to make an accurate diagnosis or establish fibrosis stage, but this limitation was overcome with improved techniques.^15,16 Transjugular liver biopsy has the additional advantage of enabling one to measure the hepatic venous pressure gradient, which also has prognostic significance; a gradient greater than 10 mm Hg is associated with worse prognosis.¹⁷

Disadvantages of biopsy: Complications, sampling errors

Liver biopsy has disadvantages. Reported rates of complications necessitating hospitalization using the blind method were as high as 6% in the 1970s,¹⁸ dropping to 3.2% in a 1993 study.¹⁹ Bleeding remains the most worrisome complication. With the transjugular method, major and minor complication rates are less than 1% and 7%, respectively.^15,16 Complication rates with imaging-guided biopsy are also low.

Liver biopsy is also prone to sampling error. The number of portal tracts obtained in the biopsy correlates with the accuracy of fibrosis staging, and smaller samples may lead to underestimating fibrosis stage. In patients with HCV, samples more than 15 mm long led to accurate staging diagnosis in 65% of patients, and those longer than 25 mm conferred 75% accuracy.²⁰ Also, different stages can be diagnosed from samples obtained from separate locations in the liver, although rarely is the difference more than a single stage.²¹

Histologic evaluation of liver biopsies is operator-dependent. Although significant interobserver variation has been reported for degree of inflammation, there tends to be good concordance for fibrosis staging.^22,23

Several scores based on patient characteristics and laboratory values have been developed for assessing liver fibrosis and have been specifically validated for HCV infection, NAFLD, or both. They can serve as inexpensive initial screening tests for the presence or absence of advanced fibrosis.

FIB-4 index for HCV, NAFLD

The FIB-4 index predicts the presence of advanced fibrosis using, as its name indicates, a combination of 4 factors in fibrosis: age, platelet count, and the levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), according to the formula:

FIB-4 index = (age × AST [U/L]) /
(platelet count [× 10⁹/L] × √ALT [U/L]).

The index was derived from data from 832 patients co-infected with HCV and human immunodeficiency virus.²⁴ The Ishak staging system¹⁰ for fibrosis on liver biopsy was used for confirmation, with stage 4 to 6 defined as advanced fibrosis. A cutoff value of more than 3.25 had a positive predictive value of 65% for advanced fibrosis, and to exclude advanced fibrosis, a cutoff value of less than 1.45 had a negative predictive value of 90%.

The FIB-4 index has since been validated in patients with HCV infection²⁵ and NAFLD.²⁶ In a subsequent study in 142 patients with NAFLD, the FIB-4 index was more accurate in diagnosing advanced fibrosis than the other noninvasive prediction models discussed below.²⁷

NAFLD fibrosis score

The NAFLD fibrosis score, constructed and validated only in patients with biopsy-confirmed NAFLD, incorporates age, body mass index, presence of diabetes or prediabetes, albumin level, platelet count, and AST and ALT levels.

A group of 480 patients was used to construct the score, and 253 patients were used to validate it. Using the high cutoff value of 0.676, the presence of advanced fibrosis was diagnosed with a positive predictive value of 90% in the group used to construct the model (82% in the validation group). Using the low cutoff score of –1.455, advanced fibrosis could be excluded with a negative predictive value of 93% in the construction group and 88% in the validation group.²⁸ A score between the cutoff values merits liver biopsy to determine fibrosis stage. The score is more accurate in patients with diabetes.²⁹ When used by primary care physicians, the NAFLD fibrosis score is more cost-effective than transient elastography and liver biopsy for accurately predicting advanced fibrosis.³⁰

AST-to-platelet ratio index score for HCV, NAFLD

The AST-to-platelet ratio index (APRI) score was developed in 2003 using a cohort of 270 patients with HCV and liver biopsy as the standard. A cutoff value of less than or equal to 0.5 had a negative predictive value of 86% for the absence of significant fibrosis, while a score of more than 1.5 detected the presence of significant fibrosis with a positive predictive value of 88%.³¹ The APRI score was subsequently validated for NAFLD.^27,32

FibroSure uses a patented formula

FibroSure (LabCorp; labcorp.com) uses a patented mathematical formula that takes into account age, sex, and levels of gamma-glutamyl transferase, total bilirubin, haptoglobin, apolipoprotein-A, and alpha-2 macroglobulin to assess fibrosis. Developed in 2001 for use in patients with HCV infection, it was reported to have a positive predictive value of greater than 90% and a negative predictive value of 100% for clinically significant fibrosis, defined as stage 2 to 4 based on the METAVIR staging system in the prediction model.³³ The use of FibroSure in patients with HCV was subsequently validated in various meta-analyses and systematic reviews.^34,35 It is less accurate in patients with normal ALT levels.³⁶

FibroSure also has good accuracy for predicting fibrosis stage in chronic liver disease due to other causes, including NAFLD.³⁷

The prediction models discussed above use routine laboratory tests for chronic liver disease and thus are inexpensive. The high cost of additional testing needed for FibroSure, coupled with the risk of misdiagnosis, makes its cost-effectiveness questionable.³⁸

Conventional ultrasonography (with or without vascular imaging) and computed tomography can detect cirrhosis on the basis of certain imaging characteristics,^39,40 including the nodular contour of the liver, caudate lobe hypertrophy, ascites, reversal of blood flow in the portal vein, and splenomegaly. However, they cannot detect fibrosis in its early stages.

The 3 methods discussed below provide more accurate fibrosis staging by measuring the velocity of shear waves sent across hepatic tissue. Because shear-wave velocity increases with liver stiffness, the fibrosis stage can be estimated from this information.⁴¹

Transient elastography

Transient elastography uses a special ultrasound transducer. It is highly accurate for predicting advanced fibrosis for almost all causes of chronic liver disease, including HCV infection^42,43 and NAFLD.⁴⁴ The cutoff values of wave velocity to estimate fibrosis stage differ by liver disease etiology.

Transient elastography should not be used to evaluate fibrosis in patients with acute hepatitis, which transiently increases liver stiffness, resulting in a falsely high fibrosis stage diagnosis.⁴⁵ It is also not a good method for evaluating fibrosis in patients with biliary obstruction or extrahepatic venous congestion. Because liver stiffness can increase after eating,⁴⁶ the test should be done under fasting conditions.

A significant limitation of transient elastography has been its poor accuracy in patients with obesity.⁴⁷ This has been largely overcome with the use of a more powerful (XL) probe but is still a limitation for those with morbid obesity.⁴⁸ Because many patients with NAFLD are obese, this limitation can be significant.

Transient elastography has gained popularity for evaluating fibrosis in patients with chronic liver disease for multiple reasons: it is cost-effective and results are highly reproducible, with low variation in results among different observers and in individual observers.⁴⁹ Combined with a platelet count, it can also be used to detect the development of clinically significant portal hypertension in patients with cirrhosis, thus determining the need to screen for esophageal varices using endoscopy.⁵⁰ Screening endoscopy can be avoided in patients whose liver stiffness remains below 20 kPa or whose platelet count is above 150 × 10⁹/L.

Acoustic radiation force imaging

Unlike transient elastography, which requires a separate transducer probe to assess shear- wave velocity, acoustic radiation force imaging uses the same transducer for both this function and imaging. Different image modes are available when testing for liver stiffness, so a region of interest that is optimal for avoiding vascular structures or masses can be selected, increasing accuracy.⁵¹

Acoustic radiation force imaging has been tested in different causes of chronic liver disease, including HCV and NAFLD,⁵² with accuracy similar to that of transient elastography.⁵³ For overweight and obese patients, acoustic radiation force imaging is more accurate than transient elastography using the XL probe.⁵⁴ However, this method is still new, and we need more data to support using one method over the other.

Magnetic resonance elastography

Magnetic resonance elastography uses a special transducer placed under the rib cage to transmit shear waves concurrently with magnetic resonance imaging. It has been tested in patients with HCV and NAFLD and has been found to have better diagnostic accuracy than transient elastography and acoustic radiation force imaging.^55,56 Patients must be fasting for better diagnostic accuracy⁵⁷ and must hold their breath while elastography is performed. The need for breath-holding and the high cost limit the use of this method for assessing fibrosis.

BOTTOM LINE FOR ASSESSING FIBROSIS

Staging of liver fibrosis, important for determining prognosis in patients with chronic liver disease and for the need to start screening for complications of cirrhosis, was traditionally done only by liver biopsy. While biopsy is still the gold standard method to stage fibrosis, noninvasive methods have been developed that can also assess disease severity.

This article briefly reviews the epidemiology and physiology of chronic liver disease and the traditional role of liver biopsy. Pros and cons of alternative fibrosis assessment methods are discussed, with a focus on their utility for patients with nonalcoholic fatty liver disease (NAFLD) and hepatitis C virus (HCV) infection.

CHRONIC LIVER DISEASE: A HUGE HEALTH BURDEN

Chronic liver disease is associated with enormous health and financial costs in the United States. Its prevalence is about 15%,¹ and it is the 12th leading cause of death.² Hospital costs are estimated at about $4 billion annually.³

The most common causes of chronic liver disease are NAFLD (which may be present in up to one-third of the US population and is increasing with the epidemic of obesity), its aggressive variant, nonalcoholic steatohepatitis (NASH) (present in about 3% of the population), and HCV infection (1%).^4,5

Since direct-acting antiviral agents were introduced, HCV infection dropped from being the leading cause of liver transplant to third place.⁶ But at the same time, the number of patients on the transplant waiting list who have NASH has risen faster than for any other cause of chronic liver disease.⁷

FIBROSIS: A KEY INDICATOR OF DISEASE SEVERITY

In HCV infection, advanced fibrosis is defined as either stage 4 to 6 using the Ishak system¹⁰ or stage 3 to 4 using the Meta-analysis of Histological Data in Viral Hepatitis (METAVIR) system.¹¹

In NAFLD, advanced fibrosis is defined as stage 3 to 4 using the NASH Clinical Research Network system.¹²

Staging fibrosis is also important so that patients with cirrhosis can be identified early to begin screening for hepatocellular carcinoma and esophageal varices to reduce the risks of illness and death. In addition, insurance companies often require documentation of fibrosis stage before treating HCV with the new direct-acting antiviral agents.

LIVER BIOPSY IS STILL THE GOLD STANDARD

Although invasive, liver biopsy remains the gold standard for determining fibrosis stage. Liver biopsies were performed “blindly” (without imaging) until the 1990s, but imaging-guided biopsy using ultrasonography was then developed, which entailed less pain and lower complication and hospitalization rates. Slightly more hepatic tissue is obtained with guided liver biopsy, but the difference was deemed clinically insignificant.¹³ Concern initially arose about the added cost involved with imaging, but imaging-guided biopsy was actually found to be more cost-effective.¹⁴

In the 2000s, transjugular liver biopsy via the right internal jugular vein became available. This method was originally used primarily in patients with ascites or significant coagulopathy. At first, there were concerns about the adequacy of specimens obtained to make an accurate diagnosis or establish fibrosis stage, but this limitation was overcome with improved techniques.^15,16 Transjugular liver biopsy has the additional advantage of enabling one to measure the hepatic venous pressure gradient, which also has prognostic significance; a gradient greater than 10 mm Hg is associated with worse prognosis.¹⁷

Disadvantages of biopsy: Complications, sampling errors

Liver biopsy has disadvantages. Reported rates of complications necessitating hospitalization using the blind method were as high as 6% in the 1970s,¹⁸ dropping to 3.2% in a 1993 study.¹⁹ Bleeding remains the most worrisome complication. With the transjugular method, major and minor complication rates are less than 1% and 7%, respectively.^15,16 Complication rates with imaging-guided biopsy are also low.

Liver biopsy is also prone to sampling error. The number of portal tracts obtained in the biopsy correlates with the accuracy of fibrosis staging, and smaller samples may lead to underestimating fibrosis stage. In patients with HCV, samples more than 15 mm long led to accurate staging diagnosis in 65% of patients, and those longer than 25 mm conferred 75% accuracy.²⁰ Also, different stages can be diagnosed from samples obtained from separate locations in the liver, although rarely is the difference more than a single stage.²¹

Histologic evaluation of liver biopsies is operator-dependent. Although significant interobserver variation has been reported for degree of inflammation, there tends to be good concordance for fibrosis staging.^22,23

Several scores based on patient characteristics and laboratory values have been developed for assessing liver fibrosis and have been specifically validated for HCV infection, NAFLD, or both. They can serve as inexpensive initial screening tests for the presence or absence of advanced fibrosis.

FIB-4 index for HCV, NAFLD

The FIB-4 index predicts the presence of advanced fibrosis using, as its name indicates, a combination of 4 factors in fibrosis: age, platelet count, and the levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), according to the formula:

FIB-4 index = (age × AST [U/L]) /
(platelet count [× 10⁹/L] × √ALT [U/L]).

The index was derived from data from 832 patients co-infected with HCV and human immunodeficiency virus.²⁴ The Ishak staging system¹⁰ for fibrosis on liver biopsy was used for confirmation, with stage 4 to 6 defined as advanced fibrosis. A cutoff value of more than 3.25 had a positive predictive value of 65% for advanced fibrosis, and to exclude advanced fibrosis, a cutoff value of less than 1.45 had a negative predictive value of 90%.

The FIB-4 index has since been validated in patients with HCV infection²⁵ and NAFLD.²⁶ In a subsequent study in 142 patients with NAFLD, the FIB-4 index was more accurate in diagnosing advanced fibrosis than the other noninvasive prediction models discussed below.²⁷

NAFLD fibrosis score

The NAFLD fibrosis score, constructed and validated only in patients with biopsy-confirmed NAFLD, incorporates age, body mass index, presence of diabetes or prediabetes, albumin level, platelet count, and AST and ALT levels.

A group of 480 patients was used to construct the score, and 253 patients were used to validate it. Using the high cutoff value of 0.676, the presence of advanced fibrosis was diagnosed with a positive predictive value of 90% in the group used to construct the model (82% in the validation group). Using the low cutoff score of –1.455, advanced fibrosis could be excluded with a negative predictive value of 93% in the construction group and 88% in the validation group.²⁸ A score between the cutoff values merits liver biopsy to determine fibrosis stage. The score is more accurate in patients with diabetes.²⁹ When used by primary care physicians, the NAFLD fibrosis score is more cost-effective than transient elastography and liver biopsy for accurately predicting advanced fibrosis.³⁰

AST-to-platelet ratio index score for HCV, NAFLD

The AST-to-platelet ratio index (APRI) score was developed in 2003 using a cohort of 270 patients with HCV and liver biopsy as the standard. A cutoff value of less than or equal to 0.5 had a negative predictive value of 86% for the absence of significant fibrosis, while a score of more than 1.5 detected the presence of significant fibrosis with a positive predictive value of 88%.³¹ The APRI score was subsequently validated for NAFLD.^27,32

FibroSure uses a patented formula

FibroSure (LabCorp; labcorp.com) uses a patented mathematical formula that takes into account age, sex, and levels of gamma-glutamyl transferase, total bilirubin, haptoglobin, apolipoprotein-A, and alpha-2 macroglobulin to assess fibrosis. Developed in 2001 for use in patients with HCV infection, it was reported to have a positive predictive value of greater than 90% and a negative predictive value of 100% for clinically significant fibrosis, defined as stage 2 to 4 based on the METAVIR staging system in the prediction model.³³ The use of FibroSure in patients with HCV was subsequently validated in various meta-analyses and systematic reviews.^34,35 It is less accurate in patients with normal ALT levels.³⁶

FibroSure also has good accuracy for predicting fibrosis stage in chronic liver disease due to other causes, including NAFLD.³⁷

The prediction models discussed above use routine laboratory tests for chronic liver disease and thus are inexpensive. The high cost of additional testing needed for FibroSure, coupled with the risk of misdiagnosis, makes its cost-effectiveness questionable.³⁸

Conventional ultrasonography (with or without vascular imaging) and computed tomography can detect cirrhosis on the basis of certain imaging characteristics,^39,40 including the nodular contour of the liver, caudate lobe hypertrophy, ascites, reversal of blood flow in the portal vein, and splenomegaly. However, they cannot detect fibrosis in its early stages.

The 3 methods discussed below provide more accurate fibrosis staging by measuring the velocity of shear waves sent across hepatic tissue. Because shear-wave velocity increases with liver stiffness, the fibrosis stage can be estimated from this information.⁴¹

Transient elastography

Transient elastography uses a special ultrasound transducer. It is highly accurate for predicting advanced fibrosis for almost all causes of chronic liver disease, including HCV infection^42,43 and NAFLD.⁴⁴ The cutoff values of wave velocity to estimate fibrosis stage differ by liver disease etiology.

Transient elastography should not be used to evaluate fibrosis in patients with acute hepatitis, which transiently increases liver stiffness, resulting in a falsely high fibrosis stage diagnosis.⁴⁵ It is also not a good method for evaluating fibrosis in patients with biliary obstruction or extrahepatic venous congestion. Because liver stiffness can increase after eating,⁴⁶ the test should be done under fasting conditions.

A significant limitation of transient elastography has been its poor accuracy in patients with obesity.⁴⁷ This has been largely overcome with the use of a more powerful (XL) probe but is still a limitation for those with morbid obesity.⁴⁸ Because many patients with NAFLD are obese, this limitation can be significant.

Transient elastography has gained popularity for evaluating fibrosis in patients with chronic liver disease for multiple reasons: it is cost-effective and results are highly reproducible, with low variation in results among different observers and in individual observers.⁴⁹ Combined with a platelet count, it can also be used to detect the development of clinically significant portal hypertension in patients with cirrhosis, thus determining the need to screen for esophageal varices using endoscopy.⁵⁰ Screening endoscopy can be avoided in patients whose liver stiffness remains below 20 kPa or whose platelet count is above 150 × 10⁹/L.

Acoustic radiation force imaging

Unlike transient elastography, which requires a separate transducer probe to assess shear- wave velocity, acoustic radiation force imaging uses the same transducer for both this function and imaging. Different image modes are available when testing for liver stiffness, so a region of interest that is optimal for avoiding vascular structures or masses can be selected, increasing accuracy.⁵¹

Acoustic radiation force imaging has been tested in different causes of chronic liver disease, including HCV and NAFLD,⁵² with accuracy similar to that of transient elastography.⁵³ For overweight and obese patients, acoustic radiation force imaging is more accurate than transient elastography using the XL probe.⁵⁴ However, this method is still new, and we need more data to support using one method over the other.

Magnetic resonance elastography

Magnetic resonance elastography uses a special transducer placed under the rib cage to transmit shear waves concurrently with magnetic resonance imaging. It has been tested in patients with HCV and NAFLD and has been found to have better diagnostic accuracy than transient elastography and acoustic radiation force imaging.^55,56 Patients must be fasting for better diagnostic accuracy⁵⁷ and must hold their breath while elastography is performed. The need for breath-holding and the high cost limit the use of this method for assessing fibrosis.

BOTTOM LINE FOR ASSESSING FIBROSIS

Staging of liver fibrosis, important for determining prognosis in patients with chronic liver disease and for the need to start screening for complications of cirrhosis, was traditionally done only by liver biopsy. While biopsy is still the gold standard method to stage fibrosis, noninvasive methods have been developed that can also assess disease severity.

This article briefly reviews the epidemiology and physiology of chronic liver disease and the traditional role of liver biopsy. Pros and cons of alternative fibrosis assessment methods are discussed, with a focus on their utility for patients with nonalcoholic fatty liver disease (NAFLD) and hepatitis C virus (HCV) infection.

CHRONIC LIVER DISEASE: A HUGE HEALTH BURDEN

Chronic liver disease is associated with enormous health and financial costs in the United States. Its prevalence is about 15%,¹ and it is the 12th leading cause of death.² Hospital costs are estimated at about $4 billion annually.³

The most common causes of chronic liver disease are NAFLD (which may be present in up to one-third of the US population and is increasing with the epidemic of obesity), its aggressive variant, nonalcoholic steatohepatitis (NASH) (present in about 3% of the population), and HCV infection (1%).^4,5

Since direct-acting antiviral agents were introduced, HCV infection dropped from being the leading cause of liver transplant to third place.⁶ But at the same time, the number of patients on the transplant waiting list who have NASH has risen faster than for any other cause of chronic liver disease.⁷

FIBROSIS: A KEY INDICATOR OF DISEASE SEVERITY

In HCV infection, advanced fibrosis is defined as either stage 4 to 6 using the Ishak system¹⁰ or stage 3 to 4 using the Meta-analysis of Histological Data in Viral Hepatitis (METAVIR) system.¹¹

In NAFLD, advanced fibrosis is defined as stage 3 to 4 using the NASH Clinical Research Network system.¹²

Staging fibrosis is also important so that patients with cirrhosis can be identified early to begin screening for hepatocellular carcinoma and esophageal varices to reduce the risks of illness and death. In addition, insurance companies often require documentation of fibrosis stage before treating HCV with the new direct-acting antiviral agents.

LIVER BIOPSY IS STILL THE GOLD STANDARD

Although invasive, liver biopsy remains the gold standard for determining fibrosis stage. Liver biopsies were performed “blindly” (without imaging) until the 1990s, but imaging-guided biopsy using ultrasonography was then developed, which entailed less pain and lower complication and hospitalization rates. Slightly more hepatic tissue is obtained with guided liver biopsy, but the difference was deemed clinically insignificant.¹³ Concern initially arose about the added cost involved with imaging, but imaging-guided biopsy was actually found to be more cost-effective.¹⁴

In the 2000s, transjugular liver biopsy via the right internal jugular vein became available. This method was originally used primarily in patients with ascites or significant coagulopathy. At first, there were concerns about the adequacy of specimens obtained to make an accurate diagnosis or establish fibrosis stage, but this limitation was overcome with improved techniques.^15,16 Transjugular liver biopsy has the additional advantage of enabling one to measure the hepatic venous pressure gradient, which also has prognostic significance; a gradient greater than 10 mm Hg is associated with worse prognosis.¹⁷

Disadvantages of biopsy: Complications, sampling errors

Liver biopsy has disadvantages. Reported rates of complications necessitating hospitalization using the blind method were as high as 6% in the 1970s,¹⁸ dropping to 3.2% in a 1993 study.¹⁹ Bleeding remains the most worrisome complication. With the transjugular method, major and minor complication rates are less than 1% and 7%, respectively.^15,16 Complication rates with imaging-guided biopsy are also low.

Liver biopsy is also prone to sampling error. The number of portal tracts obtained in the biopsy correlates with the accuracy of fibrosis staging, and smaller samples may lead to underestimating fibrosis stage. In patients with HCV, samples more than 15 mm long led to accurate staging diagnosis in 65% of patients, and those longer than 25 mm conferred 75% accuracy.²⁰ Also, different stages can be diagnosed from samples obtained from separate locations in the liver, although rarely is the difference more than a single stage.²¹

Histologic evaluation of liver biopsies is operator-dependent. Although significant interobserver variation has been reported for degree of inflammation, there tends to be good concordance for fibrosis staging.^22,23

Several scores based on patient characteristics and laboratory values have been developed for assessing liver fibrosis and have been specifically validated for HCV infection, NAFLD, or both. They can serve as inexpensive initial screening tests for the presence or absence of advanced fibrosis.

FIB-4 index for HCV, NAFLD

The FIB-4 index predicts the presence of advanced fibrosis using, as its name indicates, a combination of 4 factors in fibrosis: age, platelet count, and the levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), according to the formula:

FIB-4 index = (age × AST [U/L]) /
(platelet count [× 10⁹/L] × √ALT [U/L]).

The index was derived from data from 832 patients co-infected with HCV and human immunodeficiency virus.²⁴ The Ishak staging system¹⁰ for fibrosis on liver biopsy was used for confirmation, with stage 4 to 6 defined as advanced fibrosis. A cutoff value of more than 3.25 had a positive predictive value of 65% for advanced fibrosis, and to exclude advanced fibrosis, a cutoff value of less than 1.45 had a negative predictive value of 90%.

The FIB-4 index has since been validated in patients with HCV infection²⁵ and NAFLD.²⁶ In a subsequent study in 142 patients with NAFLD, the FIB-4 index was more accurate in diagnosing advanced fibrosis than the other noninvasive prediction models discussed below.²⁷

NAFLD fibrosis score

The NAFLD fibrosis score, constructed and validated only in patients with biopsy-confirmed NAFLD, incorporates age, body mass index, presence of diabetes or prediabetes, albumin level, platelet count, and AST and ALT levels.

A group of 480 patients was used to construct the score, and 253 patients were used to validate it. Using the high cutoff value of 0.676, the presence of advanced fibrosis was diagnosed with a positive predictive value of 90% in the group used to construct the model (82% in the validation group). Using the low cutoff score of –1.455, advanced fibrosis could be excluded with a negative predictive value of 93% in the construction group and 88% in the validation group.²⁸ A score between the cutoff values merits liver biopsy to determine fibrosis stage. The score is more accurate in patients with diabetes.²⁹ When used by primary care physicians, the NAFLD fibrosis score is more cost-effective than transient elastography and liver biopsy for accurately predicting advanced fibrosis.³⁰

AST-to-platelet ratio index score for HCV, NAFLD

The AST-to-platelet ratio index (APRI) score was developed in 2003 using a cohort of 270 patients with HCV and liver biopsy as the standard. A cutoff value of less than or equal to 0.5 had a negative predictive value of 86% for the absence of significant fibrosis, while a score of more than 1.5 detected the presence of significant fibrosis with a positive predictive value of 88%.³¹ The APRI score was subsequently validated for NAFLD.^27,32

FibroSure uses a patented formula

FibroSure (LabCorp; labcorp.com) uses a patented mathematical formula that takes into account age, sex, and levels of gamma-glutamyl transferase, total bilirubin, haptoglobin, apolipoprotein-A, and alpha-2 macroglobulin to assess fibrosis. Developed in 2001 for use in patients with HCV infection, it was reported to have a positive predictive value of greater than 90% and a negative predictive value of 100% for clinically significant fibrosis, defined as stage 2 to 4 based on the METAVIR staging system in the prediction model.³³ The use of FibroSure in patients with HCV was subsequently validated in various meta-analyses and systematic reviews.^34,35 It is less accurate in patients with normal ALT levels.³⁶

FibroSure also has good accuracy for predicting fibrosis stage in chronic liver disease due to other causes, including NAFLD.³⁷

The prediction models discussed above use routine laboratory tests for chronic liver disease and thus are inexpensive. The high cost of additional testing needed for FibroSure, coupled with the risk of misdiagnosis, makes its cost-effectiveness questionable.³⁸

Conventional ultrasonography (with or without vascular imaging) and computed tomography can detect cirrhosis on the basis of certain imaging characteristics,^39,40 including the nodular contour of the liver, caudate lobe hypertrophy, ascites, reversal of blood flow in the portal vein, and splenomegaly. However, they cannot detect fibrosis in its early stages.

The 3 methods discussed below provide more accurate fibrosis staging by measuring the velocity of shear waves sent across hepatic tissue. Because shear-wave velocity increases with liver stiffness, the fibrosis stage can be estimated from this information.⁴¹

Transient elastography

Transient elastography uses a special ultrasound transducer. It is highly accurate for predicting advanced fibrosis for almost all causes of chronic liver disease, including HCV infection^42,43 and NAFLD.⁴⁴ The cutoff values of wave velocity to estimate fibrosis stage differ by liver disease etiology.

Transient elastography should not be used to evaluate fibrosis in patients with acute hepatitis, which transiently increases liver stiffness, resulting in a falsely high fibrosis stage diagnosis.⁴⁵ It is also not a good method for evaluating fibrosis in patients with biliary obstruction or extrahepatic venous congestion. Because liver stiffness can increase after eating,⁴⁶ the test should be done under fasting conditions.

A significant limitation of transient elastography has been its poor accuracy in patients with obesity.⁴⁷ This has been largely overcome with the use of a more powerful (XL) probe but is still a limitation for those with morbid obesity.⁴⁸ Because many patients with NAFLD are obese, this limitation can be significant.

Transient elastography has gained popularity for evaluating fibrosis in patients with chronic liver disease for multiple reasons: it is cost-effective and results are highly reproducible, with low variation in results among different observers and in individual observers.⁴⁹ Combined with a platelet count, it can also be used to detect the development of clinically significant portal hypertension in patients with cirrhosis, thus determining the need to screen for esophageal varices using endoscopy.⁵⁰ Screening endoscopy can be avoided in patients whose liver stiffness remains below 20 kPa or whose platelet count is above 150 × 10⁹/L.

Acoustic radiation force imaging

Unlike transient elastography, which requires a separate transducer probe to assess shear- wave velocity, acoustic radiation force imaging uses the same transducer for both this function and imaging. Different image modes are available when testing for liver stiffness, so a region of interest that is optimal for avoiding vascular structures or masses can be selected, increasing accuracy.⁵¹

Acoustic radiation force imaging has been tested in different causes of chronic liver disease, including HCV and NAFLD,⁵² with accuracy similar to that of transient elastography.⁵³ For overweight and obese patients, acoustic radiation force imaging is more accurate than transient elastography using the XL probe.⁵⁴ However, this method is still new, and we need more data to support using one method over the other.

Magnetic resonance elastography

Magnetic resonance elastography uses a special transducer placed under the rib cage to transmit shear waves concurrently with magnetic resonance imaging. It has been tested in patients with HCV and NAFLD and has been found to have better diagnostic accuracy than transient elastography and acoustic radiation force imaging.^55,56 Patients must be fasting for better diagnostic accuracy⁵⁷ and must hold their breath while elastography is performed. The need for breath-holding and the high cost limit the use of this method for assessing fibrosis.

BOTTOM LINE FOR ASSESSING FIBROSIS

In Reply: We thank Dr. Homler for bringing hepatitis D as a potential cause of acute liver failure to our attention.

Hepatitis D virus, first described in the 1970s, requires the hepatitis B surface antigen (HBsAg) capsid to enter the hepatocyte and, thus, can only cause liver injury when the patient is also infected simultaneously with hepatitis B virus.¹ Hepatitis D virus can cause either coinfection (presence of immunoglobulin M anti-HB core antibody with or without HBsAg) or superinfection (presence of HBsAg without immunoglobulin M anti-HB core antibody) with hepatitis B virus. In India, coinfection has been reported to be the cause of acute liver failure in about 4% of all patients, and superinfection in 4.5%.²

While simultaneous treatment for hepatitis D and B viruses with pegylated interferon and any of the agents used for treatment of hepatitis B has been successful in treating chronic hepatitis, it has not been proven useful in patients with acute liver failure, and liver transplant remains the only treatment option.³

In Reply: We thank Dr. Homler for bringing hepatitis D as a potential cause of acute liver failure to our attention.

Hepatitis D virus, first described in the 1970s, requires the hepatitis B surface antigen (HBsAg) capsid to enter the hepatocyte and, thus, can only cause liver injury when the patient is also infected simultaneously with hepatitis B virus.¹ Hepatitis D virus can cause either coinfection (presence of immunoglobulin M anti-HB core antibody with or without HBsAg) or superinfection (presence of HBsAg without immunoglobulin M anti-HB core antibody) with hepatitis B virus. In India, coinfection has been reported to be the cause of acute liver failure in about 4% of all patients, and superinfection in 4.5%.²

While simultaneous treatment for hepatitis D and B viruses with pegylated interferon and any of the agents used for treatment of hepatitis B has been successful in treating chronic hepatitis, it has not been proven useful in patients with acute liver failure, and liver transplant remains the only treatment option.³

In Reply: We thank Dr. Homler for bringing hepatitis D as a potential cause of acute liver failure to our attention.

Hepatitis D virus, first described in the 1970s, requires the hepatitis B surface antigen (HBsAg) capsid to enter the hepatocyte and, thus, can only cause liver injury when the patient is also infected simultaneously with hepatitis B virus.¹ Hepatitis D virus can cause either coinfection (presence of immunoglobulin M anti-HB core antibody with or without HBsAg) or superinfection (presence of HBsAg without immunoglobulin M anti-HB core antibody) with hepatitis B virus. In India, coinfection has been reported to be the cause of acute liver failure in about 4% of all patients, and superinfection in 4.5%.²

While simultaneous treatment for hepatitis D and B viruses with pegylated interferon and any of the agents used for treatment of hepatitis B has been successful in treating chronic hepatitis, it has not been proven useful in patients with acute liver failure, and liver transplant remains the only treatment option.³

When the liver fails, it usually fails gradually. The sudden (acute) onset of liver failure, while less common, demands prompt management, with transfer to an intensive care unit, specific treatment depending on the cause, and consideration of liver transplant, without which the mortality rate is high.

This article reviews the definition, epidemiology, etiology, and management of acute liver failure.

DEFINITIONS

Acute liver failure is defined as a syndrome of acute hepatitis with evidence of abnormal coagulation (eg, an international normalized ratio > 1.5) complicated by the development of mental alteration (encephalopathy) within 26 weeks of the onset of illness in a patient without a history of liver disease.¹ In general, patients have no evidence of underlying chronic liver disease, but there are exceptions; patients with Wilson disease, vertically acquired hepatitis B virus infection, or autoimmune hepatitis can present with acute liver failure superimposed on chronic liver disease or even cirrhosis.

The term acute liver failure has replaced older terms such as fulminant hepatic failure, hyperacute liver failure, and subacute liver failure, which were used for prognostic purposes. Patients with hyperacute liver failure (defined as development of encephalopathy within 7 days of onset of illness) generally have a good prognosis with medical management, whereas those with subacute liver failure (defined as development of encephalopathy within 5 to 26 weeks of onset of illness) have a poor prognosis without liver transplant.^2,3

NEARLY 2,000 CASES A YEAR

There are nearly 2,000 cases of acute liver failure each year in the United States, and it accounts for 6% of all deaths due to liver disease.⁴ It is more common in women than in men, and more common in white people than in other races. The peak incidence is at a fairly young age, ie, 35 to 45 years.

CAUSES

The most common cause of acute liver failure in the United States and other Western countries is acetaminophen toxicity, followed by viral hepatitis. In contrast, viral hepatitis is the most common cause in developing countries.⁵

Acetaminophen toxicity

Patients with acetaminophen-induced liver failure tend to be younger than other patients with acute liver failure.¹ Nearly half of them present after intentionally taking a single large dose, while the rest present with unintentional toxicity while taking acetaminophen for pain relief on a long-term basis and ingesting more than the recommended dose.⁶

After ingestion, 52% to 57% of acetaminophen is converted to glucuronide conjugates, and 30% to 44% is converted to sulfate conjugates. These compounds are nontoxic, water-soluble, and rapidly excreted in the urine.

However, about 5% to 10% of ingested acetaminophen is shunted to the cytochrome P450 system. P450 2E1 is the main isoenzyme involved in acetaminophen metabolism, but 1A2, 3A4, and 2A6 also contribute.^7,8 P450 2E1 is the same isoenzyme responsible for ethanol metabolism and is inducible. Thus, regular alcohol consumption can increase P450 2E1 activity, setting the stage under certain circumstances for increased acetaminophen metabolism through this pathway.

Reprinted from Schilling A, Corey R, Leonard M, Eghtesad B. Acetaminophen: old drug, new warnings. Cleve Clin J Med 2010; 77:19–27.

Metabolism of acetaminophen through the cytochrome P450 pathway results in production of N-acetyl-p-benzoquinone imine (NAPQI), the compound that damages the liver. NAPQI is rendered nontoxic by binding to glutathione, forming NAPQI-glutathione adducts. Glutathione capacity is limited, however. With too much acetaminophen, glutathione becomes depleted and NAPQI accumulates, binds with proteins to form adducts, and leads to necrosis of hepatocytes (Figure 1).^9,10

Acetylcysteine, used in treating acetaminophen toxicity, is a substrate for glutathione synthesis and ultimately increases the amount of glutathione available to bind NAPQI and prevent damage to hepatocytes.¹¹

Acetaminophen is a dose-related toxin. Most ingestions leading to acute liver failure exceed 10 g/day (> 150 mg/kg/day). Moderate chronic ingestion, eg, 4 g/day, usually leads to transient mild elevation of liver enzymes in healthy individuals¹² but can in rare cases cause acute liver failure.¹³

Whitcomb and Block¹⁴ retrospectively identified 49 patients who presented with acetaminophen-induced hepatotoxicity in 1987 through 1993; 21 (43%) had been taking acetaminophen for therapeutic purposes. All 49 patients took more than the recommended limit of 4 g/day, many of them while fasting and some while using alcohol. Acute liver failure was seen with ingestion of more than 12 g/day—or more than 10 g/day in alcohol users. The authors attributed the increased risk to activation of cytochrome P450 2E1 by alcohol and depletion of glutathione stores by starvation or alcohol abuse. 

Advice to patients taking acetaminophen is given in Table 1.

Other drugs and supplements

A number of other drugs and herbal supplements can also cause acute liver failure (Table 2), the most common being antimicrobial and antiepileptic drugs.¹⁵ Of the antimicrobials, antitubercular drugs (especially isoniazid) are believed to be the most common causes, followed by trimethoprim-sulfamethoxazole. Phenytoin is the antiepileptic drug most often implicated in acute liver failure.

Statins can also cause acute liver failure, especially when combined with other hepatotoxic agents.¹⁶

The herbal supplements and weight-loss agents Hydroxycut and Herbalife have both been reported to cause acute liver failure, with patients presenting with either the hepatocellular or the cholestatic pattern of liver injury.¹⁷ The exact chemical in these supplements that causes liver injury has not yet been determined.

The National Institutes of Health maintains a database of cases of liver failure due to medications and supplements at livertox.nih.gov. The database includes the pattern of hepatic injury, mechanism of injury, management, and outcomes.

Viral hepatitis

Hepatitis B virus is the most common viral cause of acute liver failure and is responsible for about 8% of cases.¹⁸

Patients with chronic hepatitis B virus infection—as evidenced by positive hepatitis B surface antigen—can develop acute liver failure if the infection is reactivated by the use of immunosuppressive drugs for solid-organ or bone-marrow transplant or medications such as anti-tumor necrosis agents, rituximab, or chemotherapy. These patients should be treated prophylactically with a nucleoside analogue, which should be continued for 6 months after immunosuppressive therapy is completed.

Hepatitis A virus is responsible for about 4% of cases.¹⁸

Hepatitis C virus rarely causes acute liver failure, especially in the absence of hepatitis A and hepatitis B.^3,19

Hepatitis E virus, which is endemic in areas of Asia and Africa, can cause liver disease in pregnant women and in young adults who have concomitant liver disease from another cause. It tends to cause acute liver failure more frequently in pregnant women than in the rest of the population and carries a mortality rate of more than 20% in this subgroup.

TT (transfusion-transmitted) virus was reported in the 1990s to cause acute liver failure in about 27% of patients in whom no other cause could be found.²⁰

Other rare viral causes of acute liver failure include Epstein-Barr virus, cytomegalovirus, and herpes simplex virus types 1, 2, and 6.

Other causes

Other causes of acute liver failure include ischemic hepatitis, autoimmune hepatitis, Wilson disease, Budd-Chiari syndrome, and HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome.

MANY PATIENTS NEED LIVER TRANSPLANT

Many patients with acute liver failure ultimately require orthotopic liver transplant,²¹ especially if they present with severe encephalopathy. Other aspects of treatment vary according to the cause of liver failure (Table 3).

SPECIFIC MANAGEMENT

Management of acetaminophen toxicity

If the time of ingestion is known, checking the acetaminophen level can help determine the cause of acute liver failure and also predict the risk of hepatotoxicity, based on the work of Rumack and Matthew.²² Calculators are available, eg, http://reference.medscape.com/calculator/acetaminophen-toxicity.

If a patient presents with acute liver failure several days after ingesting acetaminophen, the level can be in the nontoxic range, however. In this scenario, measuring acetaminophen-protein adducts can help establish acetaminophen toxicity as the cause, as the adducts last longer in the serum and provide 100% sensitivity and specificity.²³ While most laboratories can rapidly measure acetaminophen levels, only a few can measure acetaminophen-protein adducts, and thus this test is not used clinically.

Acetylcysteine is the main drug used for acetaminophen toxicity. Ideally, it should be given within 8 hours of acetaminophen ingestion, but giving it later is also useful.¹

Acetylcysteine is available in oral and intravenous forms, the latter for patients who have encephalopathy or cannot tolerate oral intake due to repeated episodes of vomiting.^24,25 The oral form is much less costly and is thus preferred over intravenous acetylcysteine in patients who can tolerate oral intake. Intravenous acetylcysteine should be given in a loading dose of 150 mg/kg in 5% dextrose over 15 minutes, followed by a maintenance dose of 50 mg/kg over 4 hours and then 100 mg/kg given over 16 hours.¹ No dose adjustment is needed in patients who have renal toxicity (acetaminophen can also be toxic to the kidneys).

Most patients with acetaminophen-induced liver failure survive with medical management alone and do not need a liver transplant.^3,26 Cirrhosis does not occur in these patients.

Management of viral acute liver failure

When patients present with acute liver failure, it is necessary to look for a viral cause by serologic testing, including hepatitis A virus IgM antibody, hepatitis B surface antigen, and hepatitis B core IgM antibody.

Hepatitis B can become reactivated in immunocompromised patients, and therefore the hepatitis B virus DNA level should be checked. Detection of hepatitis B virus DNA in a patient previously known to have undetectable hepatitis B virus DNA confirms hepatitis B reactivation.

Patients with hepatitis B-induced acute liver failure should be treated with entecavir or tenofovir. Although this treatment may not change the course of acute liver failure or accelerate the recovery, it can prevent reinfection in the transplanted liver if liver transplant becomes indicated.^27–29

Herpes simplex virus should be suspected in patients presenting with anicteric hepatitis with fever. Polymerase chain reaction testing for herpes simplex virus should be done,³⁰ and if positive, patients should be given intravenous acyclovir.³¹ Despite treatment, herpes simplex virus disease is associated with a very poor prognosis without liver transplant.

Autoimmune hepatitis

The autoantibodies usually seen in autoimmune hepatitis are antinuclear antibody, antismooth muscle antibody, and anti-liver-kidney microsomal antibody, and patients need to be tested for them.

The diagnosis of autoimmune hepatitis can be challenging, as these autoimmune markers can be negative in 5% of patients. Liver biopsy becomes essential to establish the diagnosis in that setting.³²

Guidelines advise starting prednisone 40 to 60 mg/day and placing the patient on the liver transplant list.¹

Wilson disease

Although it is an uncommon cause of liver failure, Wilson disease needs special attention because it has a poor prognosis. The mortality rate in acute liver failure from Wilson disease reaches 100% without liver transplant.

Wilson disease is caused by a genetic defect that allows copper to accumulate in the liver and other organs. However, diagnosing Wilson disease as the cause of acute liver failure can be challenging because elevated serum and urine copper levels are not specific to Wilson disease and can be seen in patients with acute liver failure from any cause. In addition, the ceruloplasmin level is usually normal or high because it is an acute-phase reactant. Accumulation of copper in the liver parenchyma is usually patchy; therefore, qualitative copper staining on random liver biopsy samples provides low diagnostic yield. Quantitative copper on liver biopsy is the gold standard test to establish the diagnosis, but the test is time-consuming. Kayser-Fleischer rings around the iris are considered pathognomic for Wilson disease when seen with acute liver failure, but they are seen in only about 50% of patients.³³

A unique feature of acute Wilson disease is that most patients have very high bilirubin levels and low alkaline phosphatase levels. An alkaline phosphatase-to-bilirubin ratio less than 2 in patients with acute liver failure is highly suggestive of Wilson disease.³⁴

Another clue to the diagnosis is that patients with Wilson disease tend to develop Coombs-negative hemolytic anemia, which leads to a disproportionate elevation in aminotransferase levels, with aspartate aminotransferase being higher than alanine aminotransferase.

Once Wilson disease is suspected, the patient should be listed for liver transplant because death is almost certain without it. For patients awaiting liver transplant, the American Association for the Study of Liver Diseases guidelines recommend certain measures to lower the serum copper level such as albumin dialysis, continuous hemofiltration, plasmapheresis, and plasma exchange,¹ but the evidence supporting their use is limited.

NONSPECIFIC MANAGEMENT

Acute liver failure can affect a number of organs and systems in addition to the liver (Figure 2).

General considerations

Because their condition can rapidly deteriorate, patients with acute liver failure are best managed in intensive care.

Patients who present to a center that does not have the facilities for liver transplant should be transferred to a transplant center as soon as possible, preferably by air. If the patient may not be able to protect the airway, endotracheal intubation should be performed before transfer.

The major causes of death in patients with acute liver failure are cerebral edema and infection. Gastrointestinal bleeding was a major cause of death in the past, but with prophylactic use of histamine H₂ receptor blockers and proton pump inhibitors, the incidence of gastrointestinal bleeding has been significantly reduced.

Although initially used only in patients with acetaminophen-induced liver failure, acetylcysteine has also shown benefit in patients with acute liver failure from other causes. In patients with grade 1 or 2 encephalopathy on a scale of 0 (minimal) to 4 (comatose), the transplant-free survival rate is higher when acetylcysteine is given compared with placebo, but this benefit does not extend to patients with a higher grade of encephalopathy.³⁵

Cerebral edema and intracranial hypertension

Cerebral edema is the leading cause of death in patients with acute liver failure, and it develops in nearly 40% of patients.³⁶

The mechanism by which cerebral edema develops is not well understood. Some have proposed that ammonia is converted to glutamine, which causes cerebral edema either directly by its osmotic effect^37,38 or indirectly by decreasing other osmolytes, thereby promoting water retention.³⁹

Cerebral edema leads to intracranial hypertension, which can ultimately cause cerebral herniation and death. Because of the high mortality rate associated with cerebral edema, invasive devices were extensively used in the past to monitor intracranial pressure. However, in light of known complications of these devices, including bleeding,⁴⁰ and lack of evidence of long-term benefit in terms of mortality rates, their use has come under debate.

Treatments. Many treatments are available for cerebral edema and intracranial hypertension. The first step is to elevate the head of the bed about 30 degrees. In addition, hyponatremia should be corrected, as it can worsen cerebral edema.⁴¹ If patients are intubated, maintaining a hypercapneic state is advisable to decrease the intracranial pressure.

Of the two pharmacologic options, mannitol is more often used.⁴² It is given as a bolus dose of 0.5 to 1 g/kg intravenously if the serum osmolality is less than 320 mOsm/L.¹ Given the risk of fluid overload with mannitol, caution must be exercised in patients with renal dysfunction. The other pharmacologic option is 3% hypertonic saline.

Therapeutic hypothermia is a newer treatment for cerebral edema. Lowering the body temperature to 32 to 33°C (89.6 to 91.4°F) using cooling blankets decreases intracranial pressure and cerebral blood flow and improves the cerebral perfusion pressure.⁴³ With this treatment, patients should be closely monitored for side effects of infection, coagulopathy, and cardiac arrythmias.¹

l-ornithine l-aspartate was successfully used to prevent brain edema in rats, but in humans, no benefit was seen compared with placebo.^44,45 The underlying basis for this experimental treatment is that supplemental ornithine and aspartate should increase glutamate synthesis, which should increase the activity of enzyme glutamine synthetase in skeletal muscles. With the increase in enzyme activity, conversion of ammonia to glutamine should increase, thereby decreasing ammonia circulation and thus decreasing cerebral edema.

Patients with cerebral edema have a high incidence of seizures, but prophylactic antiseizure medications such as phenytoin have not been proven to be beneficial.⁴⁶

Infection

Nearly 80% of patients with acute liver failure develop an infectious complication, which can be attributed to a state of immunodeficiency.⁴⁷

The respiratory and urinary tracts are the most common sources of infection.⁴⁸ In patients with bacteremia, Enterococcus species and coagulase-negative Staphylococcus species⁴⁹ are the commonly isolated organisms. Also, in patients with acute liver failure, fungal infections account for 30% of all infections.⁵⁰

Infected patients often develop worsening of their encephalopathy⁵¹ without fever or elevated white blood cell count.^49,52 Thus, in any patient in whom encephalopathy is worsening, an evaluation must be done to rule out infection. In these patients, systemic inflammatory response syndrome is an independent risk factor for death.⁵³

Despite the high mortality rate with infection, whether using antibiotics prophylactically in acute liver failure is beneficial is controversial.^54,55

Gastrointestinal bleeding

The current prevalence of upper gastrointestinal bleeding in acute liver failure patients is about 1.5%.⁵⁶ Coagulopathy and endotracheal intubation are the main risk factors for upper gastrointestinal bleeding in these patients.⁵⁷ The most common source of bleeding is stress ulcers in the stomach. The ulcers develop from a combination of factors, including decreased blood flow to the mucosa causing ischemia and hypoperfusion-reperfusion injury.

Pharmacologic inhibition of gastric acid secretion has been shown to reduce upper gastrointestinal bleeding in acute liver failure. A histamine H₂ receptor blocker or proton pump inhibitor should be given to prevent gastrointestinal bleeding in patients with acute liver failure.^1,58

EXPERIMENTAL TREATMENTS

Artificial liver support systems

Membranes and dialysate solutions have been developed to remove toxic substances that are normally metabolized by the liver. Two of these—the molecular adsorbent recycling system (MARS) and the extracorporeal liver assist device (ELAD)—were developed in the late 1990s. MARS consisted of a highly permeable hollow fiber membrane mixed with albumin, and ELAD consisted of porcine hepatocytes attached to microcarriers in the extracapillary space of the hollow fiber membrane. Both systems allowed for transfer of water-soluble and protein-bound toxins in the blood across the membrane and into the dialysate.⁵⁹ The clinical benefit offered by these devices is controversial,^60–62 thus limiting their use to experimental purposes only.

Hepatocyte transplant

Use of hepatocyte transplant as a bridge to liver transplant was tested in 1970s, first in rats and later in humans.⁶³ By reducing the blood ammonia level and improving cerebral perfusion pressure and cardiac function, replacement of 1% to 2% of the total liver cell mass by transplanted hepatocytes acts as a bridge to orthotopic liver transplant.^64,65

PROGNOSIS

Different criteria have been used to identify patients with poor prognosis who may eventually need to undergo liver transplant.

The King’s College criteria system is the most commonly used for prognosis (Table 4).^37,66–69 Its main drawback is that it is applicable only in patients with encephalopathy, and when patients reach this stage, their condition often deteriorates rapidly, and they die while awaiting liver transplant.^37,66,67

The Model for End-Stage Liver Disease (MELD) score is an alternative to the King’s College criteria. A high MELD score on admission signifies advanced disease, and patients with a high MELD score tend to have a worse prognosis than those with a low score.⁶⁸

The Acute Physiology and Chronic Health Evaluation (APACHE) II score can also be used, as it is more sensitive than the King’s College criteria.⁶

The Clichy criteria^66,69 can also be used.

Liver biopsy. In addition to helping establish the cause of acute liver failure, liver biopsy can also be used as a prognostic tool. Hepatocellular necrosis greater than 70% on the biopsy predicts death with a specificity of 90% and a sensitivity of 56%.⁷⁰

Hypophosphatemia has been reported to indicate recovering liver function in patients with acute liver failure.⁷¹ As the liver regenerates, its energy requirement increases. To supply the energy, adenosine triphosphate production increases, and phosphorus shifts from the extracellular to the intracellular compartment to meet the need for extra phosphorus during this process. A serum phosphorus level of 2.9 mg/dL or higher appears to indicate a poor prognosis in patients with acute liver failure, as it signifies that adequate hepatocyte regeneration is not occurring.

When the liver fails, it usually fails gradually. The sudden (acute) onset of liver failure, while less common, demands prompt management, with transfer to an intensive care unit, specific treatment depending on the cause, and consideration of liver transplant, without which the mortality rate is high.

This article reviews the definition, epidemiology, etiology, and management of acute liver failure.

DEFINITIONS

Acute liver failure is defined as a syndrome of acute hepatitis with evidence of abnormal coagulation (eg, an international normalized ratio > 1.5) complicated by the development of mental alteration (encephalopathy) within 26 weeks of the onset of illness in a patient without a history of liver disease.¹ In general, patients have no evidence of underlying chronic liver disease, but there are exceptions; patients with Wilson disease, vertically acquired hepatitis B virus infection, or autoimmune hepatitis can present with acute liver failure superimposed on chronic liver disease or even cirrhosis.

The term acute liver failure has replaced older terms such as fulminant hepatic failure, hyperacute liver failure, and subacute liver failure, which were used for prognostic purposes. Patients with hyperacute liver failure (defined as development of encephalopathy within 7 days of onset of illness) generally have a good prognosis with medical management, whereas those with subacute liver failure (defined as development of encephalopathy within 5 to 26 weeks of onset of illness) have a poor prognosis without liver transplant.^2,3

NEARLY 2,000 CASES A YEAR

There are nearly 2,000 cases of acute liver failure each year in the United States, and it accounts for 6% of all deaths due to liver disease.⁴ It is more common in women than in men, and more common in white people than in other races. The peak incidence is at a fairly young age, ie, 35 to 45 years.

CAUSES

The most common cause of acute liver failure in the United States and other Western countries is acetaminophen toxicity, followed by viral hepatitis. In contrast, viral hepatitis is the most common cause in developing countries.⁵

Acetaminophen toxicity

Patients with acetaminophen-induced liver failure tend to be younger than other patients with acute liver failure.¹ Nearly half of them present after intentionally taking a single large dose, while the rest present with unintentional toxicity while taking acetaminophen for pain relief on a long-term basis and ingesting more than the recommended dose.⁶

After ingestion, 52% to 57% of acetaminophen is converted to glucuronide conjugates, and 30% to 44% is converted to sulfate conjugates. These compounds are nontoxic, water-soluble, and rapidly excreted in the urine.

However, about 5% to 10% of ingested acetaminophen is shunted to the cytochrome P450 system. P450 2E1 is the main isoenzyme involved in acetaminophen metabolism, but 1A2, 3A4, and 2A6 also contribute.^7,8 P450 2E1 is the same isoenzyme responsible for ethanol metabolism and is inducible. Thus, regular alcohol consumption can increase P450 2E1 activity, setting the stage under certain circumstances for increased acetaminophen metabolism through this pathway.

Reprinted from Schilling A, Corey R, Leonard M, Eghtesad B. Acetaminophen: old drug, new warnings. Cleve Clin J Med 2010; 77:19–27.

Metabolism of acetaminophen through the cytochrome P450 pathway results in production of N-acetyl-p-benzoquinone imine (NAPQI), the compound that damages the liver. NAPQI is rendered nontoxic by binding to glutathione, forming NAPQI-glutathione adducts. Glutathione capacity is limited, however. With too much acetaminophen, glutathione becomes depleted and NAPQI accumulates, binds with proteins to form adducts, and leads to necrosis of hepatocytes (Figure 1).^9,10

Acetylcysteine, used in treating acetaminophen toxicity, is a substrate for glutathione synthesis and ultimately increases the amount of glutathione available to bind NAPQI and prevent damage to hepatocytes.¹¹

Acetaminophen is a dose-related toxin. Most ingestions leading to acute liver failure exceed 10 g/day (> 150 mg/kg/day). Moderate chronic ingestion, eg, 4 g/day, usually leads to transient mild elevation of liver enzymes in healthy individuals¹² but can in rare cases cause acute liver failure.¹³

Whitcomb and Block¹⁴ retrospectively identified 49 patients who presented with acetaminophen-induced hepatotoxicity in 1987 through 1993; 21 (43%) had been taking acetaminophen for therapeutic purposes. All 49 patients took more than the recommended limit of 4 g/day, many of them while fasting and some while using alcohol. Acute liver failure was seen with ingestion of more than 12 g/day—or more than 10 g/day in alcohol users. The authors attributed the increased risk to activation of cytochrome P450 2E1 by alcohol and depletion of glutathione stores by starvation or alcohol abuse. 

Advice to patients taking acetaminophen is given in Table 1.

Other drugs and supplements

A number of other drugs and herbal supplements can also cause acute liver failure (Table 2), the most common being antimicrobial and antiepileptic drugs.¹⁵ Of the antimicrobials, antitubercular drugs (especially isoniazid) are believed to be the most common causes, followed by trimethoprim-sulfamethoxazole. Phenytoin is the antiepileptic drug most often implicated in acute liver failure.

Statins can also cause acute liver failure, especially when combined with other hepatotoxic agents.¹⁶

The herbal supplements and weight-loss agents Hydroxycut and Herbalife have both been reported to cause acute liver failure, with patients presenting with either the hepatocellular or the cholestatic pattern of liver injury.¹⁷ The exact chemical in these supplements that causes liver injury has not yet been determined.

The National Institutes of Health maintains a database of cases of liver failure due to medications and supplements at livertox.nih.gov. The database includes the pattern of hepatic injury, mechanism of injury, management, and outcomes.

Viral hepatitis

Hepatitis B virus is the most common viral cause of acute liver failure and is responsible for about 8% of cases.¹⁸

Patients with chronic hepatitis B virus infection—as evidenced by positive hepatitis B surface antigen—can develop acute liver failure if the infection is reactivated by the use of immunosuppressive drugs for solid-organ or bone-marrow transplant or medications such as anti-tumor necrosis agents, rituximab, or chemotherapy. These patients should be treated prophylactically with a nucleoside analogue, which should be continued for 6 months after immunosuppressive therapy is completed.

Hepatitis A virus is responsible for about 4% of cases.¹⁸

Hepatitis C virus rarely causes acute liver failure, especially in the absence of hepatitis A and hepatitis B.^3,19

Hepatitis E virus, which is endemic in areas of Asia and Africa, can cause liver disease in pregnant women and in young adults who have concomitant liver disease from another cause. It tends to cause acute liver failure more frequently in pregnant women than in the rest of the population and carries a mortality rate of more than 20% in this subgroup.

TT (transfusion-transmitted) virus was reported in the 1990s to cause acute liver failure in about 27% of patients in whom no other cause could be found.²⁰

Other rare viral causes of acute liver failure include Epstein-Barr virus, cytomegalovirus, and herpes simplex virus types 1, 2, and 6.

Other causes

Other causes of acute liver failure include ischemic hepatitis, autoimmune hepatitis, Wilson disease, Budd-Chiari syndrome, and HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome.

MANY PATIENTS NEED LIVER TRANSPLANT

Many patients with acute liver failure ultimately require orthotopic liver transplant,²¹ especially if they present with severe encephalopathy. Other aspects of treatment vary according to the cause of liver failure (Table 3).

SPECIFIC MANAGEMENT

Management of acetaminophen toxicity

If the time of ingestion is known, checking the acetaminophen level can help determine the cause of acute liver failure and also predict the risk of hepatotoxicity, based on the work of Rumack and Matthew.²² Calculators are available, eg, http://reference.medscape.com/calculator/acetaminophen-toxicity.

If a patient presents with acute liver failure several days after ingesting acetaminophen, the level can be in the nontoxic range, however. In this scenario, measuring acetaminophen-protein adducts can help establish acetaminophen toxicity as the cause, as the adducts last longer in the serum and provide 100% sensitivity and specificity.²³ While most laboratories can rapidly measure acetaminophen levels, only a few can measure acetaminophen-protein adducts, and thus this test is not used clinically.

Acetylcysteine is the main drug used for acetaminophen toxicity. Ideally, it should be given within 8 hours of acetaminophen ingestion, but giving it later is also useful.¹

Acetylcysteine is available in oral and intravenous forms, the latter for patients who have encephalopathy or cannot tolerate oral intake due to repeated episodes of vomiting.^24,25 The oral form is much less costly and is thus preferred over intravenous acetylcysteine in patients who can tolerate oral intake. Intravenous acetylcysteine should be given in a loading dose of 150 mg/kg in 5% dextrose over 15 minutes, followed by a maintenance dose of 50 mg/kg over 4 hours and then 100 mg/kg given over 16 hours.¹ No dose adjustment is needed in patients who have renal toxicity (acetaminophen can also be toxic to the kidneys).

Most patients with acetaminophen-induced liver failure survive with medical management alone and do not need a liver transplant.^3,26 Cirrhosis does not occur in these patients.

Management of viral acute liver failure

When patients present with acute liver failure, it is necessary to look for a viral cause by serologic testing, including hepatitis A virus IgM antibody, hepatitis B surface antigen, and hepatitis B core IgM antibody.

Hepatitis B can become reactivated in immunocompromised patients, and therefore the hepatitis B virus DNA level should be checked. Detection of hepatitis B virus DNA in a patient previously known to have undetectable hepatitis B virus DNA confirms hepatitis B reactivation.

Patients with hepatitis B-induced acute liver failure should be treated with entecavir or tenofovir. Although this treatment may not change the course of acute liver failure or accelerate the recovery, it can prevent reinfection in the transplanted liver if liver transplant becomes indicated.^27–29

Herpes simplex virus should be suspected in patients presenting with anicteric hepatitis with fever. Polymerase chain reaction testing for herpes simplex virus should be done,³⁰ and if positive, patients should be given intravenous acyclovir.³¹ Despite treatment, herpes simplex virus disease is associated with a very poor prognosis without liver transplant.

Autoimmune hepatitis

The autoantibodies usually seen in autoimmune hepatitis are antinuclear antibody, antismooth muscle antibody, and anti-liver-kidney microsomal antibody, and patients need to be tested for them.

The diagnosis of autoimmune hepatitis can be challenging, as these autoimmune markers can be negative in 5% of patients. Liver biopsy becomes essential to establish the diagnosis in that setting.³²

Guidelines advise starting prednisone 40 to 60 mg/day and placing the patient on the liver transplant list.¹

Wilson disease

Although it is an uncommon cause of liver failure, Wilson disease needs special attention because it has a poor prognosis. The mortality rate in acute liver failure from Wilson disease reaches 100% without liver transplant.

Wilson disease is caused by a genetic defect that allows copper to accumulate in the liver and other organs. However, diagnosing Wilson disease as the cause of acute liver failure can be challenging because elevated serum and urine copper levels are not specific to Wilson disease and can be seen in patients with acute liver failure from any cause. In addition, the ceruloplasmin level is usually normal or high because it is an acute-phase reactant. Accumulation of copper in the liver parenchyma is usually patchy; therefore, qualitative copper staining on random liver biopsy samples provides low diagnostic yield. Quantitative copper on liver biopsy is the gold standard test to establish the diagnosis, but the test is time-consuming. Kayser-Fleischer rings around the iris are considered pathognomic for Wilson disease when seen with acute liver failure, but they are seen in only about 50% of patients.³³

A unique feature of acute Wilson disease is that most patients have very high bilirubin levels and low alkaline phosphatase levels. An alkaline phosphatase-to-bilirubin ratio less than 2 in patients with acute liver failure is highly suggestive of Wilson disease.³⁴

Another clue to the diagnosis is that patients with Wilson disease tend to develop Coombs-negative hemolytic anemia, which leads to a disproportionate elevation in aminotransferase levels, with aspartate aminotransferase being higher than alanine aminotransferase.

Once Wilson disease is suspected, the patient should be listed for liver transplant because death is almost certain without it. For patients awaiting liver transplant, the American Association for the Study of Liver Diseases guidelines recommend certain measures to lower the serum copper level such as albumin dialysis, continuous hemofiltration, plasmapheresis, and plasma exchange,¹ but the evidence supporting their use is limited.

NONSPECIFIC MANAGEMENT

Acute liver failure can affect a number of organs and systems in addition to the liver (Figure 2).

General considerations

Because their condition can rapidly deteriorate, patients with acute liver failure are best managed in intensive care.

Patients who present to a center that does not have the facilities for liver transplant should be transferred to a transplant center as soon as possible, preferably by air. If the patient may not be able to protect the airway, endotracheal intubation should be performed before transfer.

The major causes of death in patients with acute liver failure are cerebral edema and infection. Gastrointestinal bleeding was a major cause of death in the past, but with prophylactic use of histamine H₂ receptor blockers and proton pump inhibitors, the incidence of gastrointestinal bleeding has been significantly reduced.

Although initially used only in patients with acetaminophen-induced liver failure, acetylcysteine has also shown benefit in patients with acute liver failure from other causes. In patients with grade 1 or 2 encephalopathy on a scale of 0 (minimal) to 4 (comatose), the transplant-free survival rate is higher when acetylcysteine is given compared with placebo, but this benefit does not extend to patients with a higher grade of encephalopathy.³⁵

Cerebral edema and intracranial hypertension

Cerebral edema is the leading cause of death in patients with acute liver failure, and it develops in nearly 40% of patients.³⁶

The mechanism by which cerebral edema develops is not well understood. Some have proposed that ammonia is converted to glutamine, which causes cerebral edema either directly by its osmotic effect^37,38 or indirectly by decreasing other osmolytes, thereby promoting water retention.³⁹

Cerebral edema leads to intracranial hypertension, which can ultimately cause cerebral herniation and death. Because of the high mortality rate associated with cerebral edema, invasive devices were extensively used in the past to monitor intracranial pressure. However, in light of known complications of these devices, including bleeding,⁴⁰ and lack of evidence of long-term benefit in terms of mortality rates, their use has come under debate.

Treatments. Many treatments are available for cerebral edema and intracranial hypertension. The first step is to elevate the head of the bed about 30 degrees. In addition, hyponatremia should be corrected, as it can worsen cerebral edema.⁴¹ If patients are intubated, maintaining a hypercapneic state is advisable to decrease the intracranial pressure.

Of the two pharmacologic options, mannitol is more often used.⁴² It is given as a bolus dose of 0.5 to 1 g/kg intravenously if the serum osmolality is less than 320 mOsm/L.¹ Given the risk of fluid overload with mannitol, caution must be exercised in patients with renal dysfunction. The other pharmacologic option is 3% hypertonic saline.

Therapeutic hypothermia is a newer treatment for cerebral edema. Lowering the body temperature to 32 to 33°C (89.6 to 91.4°F) using cooling blankets decreases intracranial pressure and cerebral blood flow and improves the cerebral perfusion pressure.⁴³ With this treatment, patients should be closely monitored for side effects of infection, coagulopathy, and cardiac arrythmias.¹

l-ornithine l-aspartate was successfully used to prevent brain edema in rats, but in humans, no benefit was seen compared with placebo.^44,45 The underlying basis for this experimental treatment is that supplemental ornithine and aspartate should increase glutamate synthesis, which should increase the activity of enzyme glutamine synthetase in skeletal muscles. With the increase in enzyme activity, conversion of ammonia to glutamine should increase, thereby decreasing ammonia circulation and thus decreasing cerebral edema.

Patients with cerebral edema have a high incidence of seizures, but prophylactic antiseizure medications such as phenytoin have not been proven to be beneficial.⁴⁶

Infection

Nearly 80% of patients with acute liver failure develop an infectious complication, which can be attributed to a state of immunodeficiency.⁴⁷

The respiratory and urinary tracts are the most common sources of infection.⁴⁸ In patients with bacteremia, Enterococcus species and coagulase-negative Staphylococcus species⁴⁹ are the commonly isolated organisms. Also, in patients with acute liver failure, fungal infections account for 30% of all infections.⁵⁰

Infected patients often develop worsening of their encephalopathy⁵¹ without fever or elevated white blood cell count.^49,52 Thus, in any patient in whom encephalopathy is worsening, an evaluation must be done to rule out infection. In these patients, systemic inflammatory response syndrome is an independent risk factor for death.⁵³

Despite the high mortality rate with infection, whether using antibiotics prophylactically in acute liver failure is beneficial is controversial.^54,55

Gastrointestinal bleeding

The current prevalence of upper gastrointestinal bleeding in acute liver failure patients is about 1.5%.⁵⁶ Coagulopathy and endotracheal intubation are the main risk factors for upper gastrointestinal bleeding in these patients.⁵⁷ The most common source of bleeding is stress ulcers in the stomach. The ulcers develop from a combination of factors, including decreased blood flow to the mucosa causing ischemia and hypoperfusion-reperfusion injury.

Pharmacologic inhibition of gastric acid secretion has been shown to reduce upper gastrointestinal bleeding in acute liver failure. A histamine H₂ receptor blocker or proton pump inhibitor should be given to prevent gastrointestinal bleeding in patients with acute liver failure.^1,58

EXPERIMENTAL TREATMENTS

Artificial liver support systems

Membranes and dialysate solutions have been developed to remove toxic substances that are normally metabolized by the liver. Two of these—the molecular adsorbent recycling system (MARS) and the extracorporeal liver assist device (ELAD)—were developed in the late 1990s. MARS consisted of a highly permeable hollow fiber membrane mixed with albumin, and ELAD consisted of porcine hepatocytes attached to microcarriers in the extracapillary space of the hollow fiber membrane. Both systems allowed for transfer of water-soluble and protein-bound toxins in the blood across the membrane and into the dialysate.⁵⁹ The clinical benefit offered by these devices is controversial,^60–62 thus limiting their use to experimental purposes only.

Hepatocyte transplant

Use of hepatocyte transplant as a bridge to liver transplant was tested in 1970s, first in rats and later in humans.⁶³ By reducing the blood ammonia level and improving cerebral perfusion pressure and cardiac function, replacement of 1% to 2% of the total liver cell mass by transplanted hepatocytes acts as a bridge to orthotopic liver transplant.^64,65

PROGNOSIS

Different criteria have been used to identify patients with poor prognosis who may eventually need to undergo liver transplant.

The King’s College criteria system is the most commonly used for prognosis (Table 4).^37,66–69 Its main drawback is that it is applicable only in patients with encephalopathy, and when patients reach this stage, their condition often deteriorates rapidly, and they die while awaiting liver transplant.^37,66,67

The Model for End-Stage Liver Disease (MELD) score is an alternative to the King’s College criteria. A high MELD score on admission signifies advanced disease, and patients with a high MELD score tend to have a worse prognosis than those with a low score.⁶⁸

The Acute Physiology and Chronic Health Evaluation (APACHE) II score can also be used, as it is more sensitive than the King’s College criteria.⁶

The Clichy criteria^66,69 can also be used.

Liver biopsy. In addition to helping establish the cause of acute liver failure, liver biopsy can also be used as a prognostic tool. Hepatocellular necrosis greater than 70% on the biopsy predicts death with a specificity of 90% and a sensitivity of 56%.⁷⁰

Hypophosphatemia has been reported to indicate recovering liver function in patients with acute liver failure.⁷¹ As the liver regenerates, its energy requirement increases. To supply the energy, adenosine triphosphate production increases, and phosphorus shifts from the extracellular to the intracellular compartment to meet the need for extra phosphorus during this process. A serum phosphorus level of 2.9 mg/dL or higher appears to indicate a poor prognosis in patients with acute liver failure, as it signifies that adequate hepatocyte regeneration is not occurring.

When the liver fails, it usually fails gradually. The sudden (acute) onset of liver failure, while less common, demands prompt management, with transfer to an intensive care unit, specific treatment depending on the cause, and consideration of liver transplant, without which the mortality rate is high.

This article reviews the definition, epidemiology, etiology, and management of acute liver failure.

DEFINITIONS

Acute liver failure is defined as a syndrome of acute hepatitis with evidence of abnormal coagulation (eg, an international normalized ratio > 1.5) complicated by the development of mental alteration (encephalopathy) within 26 weeks of the onset of illness in a patient without a history of liver disease.¹ In general, patients have no evidence of underlying chronic liver disease, but there are exceptions; patients with Wilson disease, vertically acquired hepatitis B virus infection, or autoimmune hepatitis can present with acute liver failure superimposed on chronic liver disease or even cirrhosis.

The term acute liver failure has replaced older terms such as fulminant hepatic failure, hyperacute liver failure, and subacute liver failure, which were used for prognostic purposes. Patients with hyperacute liver failure (defined as development of encephalopathy within 7 days of onset of illness) generally have a good prognosis with medical management, whereas those with subacute liver failure (defined as development of encephalopathy within 5 to 26 weeks of onset of illness) have a poor prognosis without liver transplant.^2,3

NEARLY 2,000 CASES A YEAR

There are nearly 2,000 cases of acute liver failure each year in the United States, and it accounts for 6% of all deaths due to liver disease.⁴ It is more common in women than in men, and more common in white people than in other races. The peak incidence is at a fairly young age, ie, 35 to 45 years.

CAUSES

The most common cause of acute liver failure in the United States and other Western countries is acetaminophen toxicity, followed by viral hepatitis. In contrast, viral hepatitis is the most common cause in developing countries.⁵

Acetaminophen toxicity

Patients with acetaminophen-induced liver failure tend to be younger than other patients with acute liver failure.¹ Nearly half of them present after intentionally taking a single large dose, while the rest present with unintentional toxicity while taking acetaminophen for pain relief on a long-term basis and ingesting more than the recommended dose.⁶

After ingestion, 52% to 57% of acetaminophen is converted to glucuronide conjugates, and 30% to 44% is converted to sulfate conjugates. These compounds are nontoxic, water-soluble, and rapidly excreted in the urine.

However, about 5% to 10% of ingested acetaminophen is shunted to the cytochrome P450 system. P450 2E1 is the main isoenzyme involved in acetaminophen metabolism, but 1A2, 3A4, and 2A6 also contribute.^7,8 P450 2E1 is the same isoenzyme responsible for ethanol metabolism and is inducible. Thus, regular alcohol consumption can increase P450 2E1 activity, setting the stage under certain circumstances for increased acetaminophen metabolism through this pathway.

Reprinted from Schilling A, Corey R, Leonard M, Eghtesad B. Acetaminophen: old drug, new warnings. Cleve Clin J Med 2010; 77:19–27.

Metabolism of acetaminophen through the cytochrome P450 pathway results in production of N-acetyl-p-benzoquinone imine (NAPQI), the compound that damages the liver. NAPQI is rendered nontoxic by binding to glutathione, forming NAPQI-glutathione adducts. Glutathione capacity is limited, however. With too much acetaminophen, glutathione becomes depleted and NAPQI accumulates, binds with proteins to form adducts, and leads to necrosis of hepatocytes (Figure 1).^9,10

Acetylcysteine, used in treating acetaminophen toxicity, is a substrate for glutathione synthesis and ultimately increases the amount of glutathione available to bind NAPQI and prevent damage to hepatocytes.¹¹

Acetaminophen is a dose-related toxin. Most ingestions leading to acute liver failure exceed 10 g/day (> 150 mg/kg/day). Moderate chronic ingestion, eg, 4 g/day, usually leads to transient mild elevation of liver enzymes in healthy individuals¹² but can in rare cases cause acute liver failure.¹³

Whitcomb and Block¹⁴ retrospectively identified 49 patients who presented with acetaminophen-induced hepatotoxicity in 1987 through 1993; 21 (43%) had been taking acetaminophen for therapeutic purposes. All 49 patients took more than the recommended limit of 4 g/day, many of them while fasting and some while using alcohol. Acute liver failure was seen with ingestion of more than 12 g/day—or more than 10 g/day in alcohol users. The authors attributed the increased risk to activation of cytochrome P450 2E1 by alcohol and depletion of glutathione stores by starvation or alcohol abuse. 

Advice to patients taking acetaminophen is given in Table 1.

Other drugs and supplements

A number of other drugs and herbal supplements can also cause acute liver failure (Table 2), the most common being antimicrobial and antiepileptic drugs.¹⁵ Of the antimicrobials, antitubercular drugs (especially isoniazid) are believed to be the most common causes, followed by trimethoprim-sulfamethoxazole. Phenytoin is the antiepileptic drug most often implicated in acute liver failure.

Statins can also cause acute liver failure, especially when combined with other hepatotoxic agents.¹⁶

The herbal supplements and weight-loss agents Hydroxycut and Herbalife have both been reported to cause acute liver failure, with patients presenting with either the hepatocellular or the cholestatic pattern of liver injury.¹⁷ The exact chemical in these supplements that causes liver injury has not yet been determined.

The National Institutes of Health maintains a database of cases of liver failure due to medications and supplements at livertox.nih.gov. The database includes the pattern of hepatic injury, mechanism of injury, management, and outcomes.

Viral hepatitis

Hepatitis B virus is the most common viral cause of acute liver failure and is responsible for about 8% of cases.¹⁸

Patients with chronic hepatitis B virus infection—as evidenced by positive hepatitis B surface antigen—can develop acute liver failure if the infection is reactivated by the use of immunosuppressive drugs for solid-organ or bone-marrow transplant or medications such as anti-tumor necrosis agents, rituximab, or chemotherapy. These patients should be treated prophylactically with a nucleoside analogue, which should be continued for 6 months after immunosuppressive therapy is completed.

Hepatitis A virus is responsible for about 4% of cases.¹⁸

Hepatitis C virus rarely causes acute liver failure, especially in the absence of hepatitis A and hepatitis B.^3,19

Hepatitis E virus, which is endemic in areas of Asia and Africa, can cause liver disease in pregnant women and in young adults who have concomitant liver disease from another cause. It tends to cause acute liver failure more frequently in pregnant women than in the rest of the population and carries a mortality rate of more than 20% in this subgroup.

TT (transfusion-transmitted) virus was reported in the 1990s to cause acute liver failure in about 27% of patients in whom no other cause could be found.²⁰

Other rare viral causes of acute liver failure include Epstein-Barr virus, cytomegalovirus, and herpes simplex virus types 1, 2, and 6.

Other causes

Other causes of acute liver failure include ischemic hepatitis, autoimmune hepatitis, Wilson disease, Budd-Chiari syndrome, and HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome.

MANY PATIENTS NEED LIVER TRANSPLANT

Many patients with acute liver failure ultimately require orthotopic liver transplant,²¹ especially if they present with severe encephalopathy. Other aspects of treatment vary according to the cause of liver failure (Table 3).

SPECIFIC MANAGEMENT

Management of acetaminophen toxicity

If the time of ingestion is known, checking the acetaminophen level can help determine the cause of acute liver failure and also predict the risk of hepatotoxicity, based on the work of Rumack and Matthew.²² Calculators are available, eg, http://reference.medscape.com/calculator/acetaminophen-toxicity.

If a patient presents with acute liver failure several days after ingesting acetaminophen, the level can be in the nontoxic range, however. In this scenario, measuring acetaminophen-protein adducts can help establish acetaminophen toxicity as the cause, as the adducts last longer in the serum and provide 100% sensitivity and specificity.²³ While most laboratories can rapidly measure acetaminophen levels, only a few can measure acetaminophen-protein adducts, and thus this test is not used clinically.

Acetylcysteine is the main drug used for acetaminophen toxicity. Ideally, it should be given within 8 hours of acetaminophen ingestion, but giving it later is also useful.¹

Acetylcysteine is available in oral and intravenous forms, the latter for patients who have encephalopathy or cannot tolerate oral intake due to repeated episodes of vomiting.^24,25 The oral form is much less costly and is thus preferred over intravenous acetylcysteine in patients who can tolerate oral intake. Intravenous acetylcysteine should be given in a loading dose of 150 mg/kg in 5% dextrose over 15 minutes, followed by a maintenance dose of 50 mg/kg over 4 hours and then 100 mg/kg given over 16 hours.¹ No dose adjustment is needed in patients who have renal toxicity (acetaminophen can also be toxic to the kidneys).

Most patients with acetaminophen-induced liver failure survive with medical management alone and do not need a liver transplant.^3,26 Cirrhosis does not occur in these patients.

Management of viral acute liver failure

When patients present with acute liver failure, it is necessary to look for a viral cause by serologic testing, including hepatitis A virus IgM antibody, hepatitis B surface antigen, and hepatitis B core IgM antibody.

Hepatitis B can become reactivated in immunocompromised patients, and therefore the hepatitis B virus DNA level should be checked. Detection of hepatitis B virus DNA in a patient previously known to have undetectable hepatitis B virus DNA confirms hepatitis B reactivation.

Patients with hepatitis B-induced acute liver failure should be treated with entecavir or tenofovir. Although this treatment may not change the course of acute liver failure or accelerate the recovery, it can prevent reinfection in the transplanted liver if liver transplant becomes indicated.^27–29

Herpes simplex virus should be suspected in patients presenting with anicteric hepatitis with fever. Polymerase chain reaction testing for herpes simplex virus should be done,³⁰ and if positive, patients should be given intravenous acyclovir.³¹ Despite treatment, herpes simplex virus disease is associated with a very poor prognosis without liver transplant.

Autoimmune hepatitis

The autoantibodies usually seen in autoimmune hepatitis are antinuclear antibody, antismooth muscle antibody, and anti-liver-kidney microsomal antibody, and patients need to be tested for them.

The diagnosis of autoimmune hepatitis can be challenging, as these autoimmune markers can be negative in 5% of patients. Liver biopsy becomes essential to establish the diagnosis in that setting.³²

Guidelines advise starting prednisone 40 to 60 mg/day and placing the patient on the liver transplant list.¹

Wilson disease

Although it is an uncommon cause of liver failure, Wilson disease needs special attention because it has a poor prognosis. The mortality rate in acute liver failure from Wilson disease reaches 100% without liver transplant.

Wilson disease is caused by a genetic defect that allows copper to accumulate in the liver and other organs. However, diagnosing Wilson disease as the cause of acute liver failure can be challenging because elevated serum and urine copper levels are not specific to Wilson disease and can be seen in patients with acute liver failure from any cause. In addition, the ceruloplasmin level is usually normal or high because it is an acute-phase reactant. Accumulation of copper in the liver parenchyma is usually patchy; therefore, qualitative copper staining on random liver biopsy samples provides low diagnostic yield. Quantitative copper on liver biopsy is the gold standard test to establish the diagnosis, but the test is time-consuming. Kayser-Fleischer rings around the iris are considered pathognomic for Wilson disease when seen with acute liver failure, but they are seen in only about 50% of patients.³³

A unique feature of acute Wilson disease is that most patients have very high bilirubin levels and low alkaline phosphatase levels. An alkaline phosphatase-to-bilirubin ratio less than 2 in patients with acute liver failure is highly suggestive of Wilson disease.³⁴

Another clue to the diagnosis is that patients with Wilson disease tend to develop Coombs-negative hemolytic anemia, which leads to a disproportionate elevation in aminotransferase levels, with aspartate aminotransferase being higher than alanine aminotransferase.

Once Wilson disease is suspected, the patient should be listed for liver transplant because death is almost certain without it. For patients awaiting liver transplant, the American Association for the Study of Liver Diseases guidelines recommend certain measures to lower the serum copper level such as albumin dialysis, continuous hemofiltration, plasmapheresis, and plasma exchange,¹ but the evidence supporting their use is limited.

NONSPECIFIC MANAGEMENT

Acute liver failure can affect a number of organs and systems in addition to the liver (Figure 2).

General considerations

Because their condition can rapidly deteriorate, patients with acute liver failure are best managed in intensive care.

Patients who present to a center that does not have the facilities for liver transplant should be transferred to a transplant center as soon as possible, preferably by air. If the patient may not be able to protect the airway, endotracheal intubation should be performed before transfer.

The major causes of death in patients with acute liver failure are cerebral edema and infection. Gastrointestinal bleeding was a major cause of death in the past, but with prophylactic use of histamine H₂ receptor blockers and proton pump inhibitors, the incidence of gastrointestinal bleeding has been significantly reduced.

Although initially used only in patients with acetaminophen-induced liver failure, acetylcysteine has also shown benefit in patients with acute liver failure from other causes. In patients with grade 1 or 2 encephalopathy on a scale of 0 (minimal) to 4 (comatose), the transplant-free survival rate is higher when acetylcysteine is given compared with placebo, but this benefit does not extend to patients with a higher grade of encephalopathy.³⁵

Cerebral edema and intracranial hypertension

Cerebral edema is the leading cause of death in patients with acute liver failure, and it develops in nearly 40% of patients.³⁶

The mechanism by which cerebral edema develops is not well understood. Some have proposed that ammonia is converted to glutamine, which causes cerebral edema either directly by its osmotic effect^37,38 or indirectly by decreasing other osmolytes, thereby promoting water retention.³⁹

Cerebral edema leads to intracranial hypertension, which can ultimately cause cerebral herniation and death. Because of the high mortality rate associated with cerebral edema, invasive devices were extensively used in the past to monitor intracranial pressure. However, in light of known complications of these devices, including bleeding,⁴⁰ and lack of evidence of long-term benefit in terms of mortality rates, their use has come under debate.

Treatments. Many treatments are available for cerebral edema and intracranial hypertension. The first step is to elevate the head of the bed about 30 degrees. In addition, hyponatremia should be corrected, as it can worsen cerebral edema.⁴¹ If patients are intubated, maintaining a hypercapneic state is advisable to decrease the intracranial pressure.

Of the two pharmacologic options, mannitol is more often used.⁴² It is given as a bolus dose of 0.5 to 1 g/kg intravenously if the serum osmolality is less than 320 mOsm/L.¹ Given the risk of fluid overload with mannitol, caution must be exercised in patients with renal dysfunction. The other pharmacologic option is 3% hypertonic saline.

Therapeutic hypothermia is a newer treatment for cerebral edema. Lowering the body temperature to 32 to 33°C (89.6 to 91.4°F) using cooling blankets decreases intracranial pressure and cerebral blood flow and improves the cerebral perfusion pressure.⁴³ With this treatment, patients should be closely monitored for side effects of infection, coagulopathy, and cardiac arrythmias.¹

l-ornithine l-aspartate was successfully used to prevent brain edema in rats, but in humans, no benefit was seen compared with placebo.^44,45 The underlying basis for this experimental treatment is that supplemental ornithine and aspartate should increase glutamate synthesis, which should increase the activity of enzyme glutamine synthetase in skeletal muscles. With the increase in enzyme activity, conversion of ammonia to glutamine should increase, thereby decreasing ammonia circulation and thus decreasing cerebral edema.

Patients with cerebral edema have a high incidence of seizures, but prophylactic antiseizure medications such as phenytoin have not been proven to be beneficial.⁴⁶

Infection

Nearly 80% of patients with acute liver failure develop an infectious complication, which can be attributed to a state of immunodeficiency.⁴⁷

The respiratory and urinary tracts are the most common sources of infection.⁴⁸ In patients with bacteremia, Enterococcus species and coagulase-negative Staphylococcus species⁴⁹ are the commonly isolated organisms. Also, in patients with acute liver failure, fungal infections account for 30% of all infections.⁵⁰

Infected patients often develop worsening of their encephalopathy⁵¹ without fever or elevated white blood cell count.^49,52 Thus, in any patient in whom encephalopathy is worsening, an evaluation must be done to rule out infection. In these patients, systemic inflammatory response syndrome is an independent risk factor for death.⁵³

Despite the high mortality rate with infection, whether using antibiotics prophylactically in acute liver failure is beneficial is controversial.^54,55

Gastrointestinal bleeding

The current prevalence of upper gastrointestinal bleeding in acute liver failure patients is about 1.5%.⁵⁶ Coagulopathy and endotracheal intubation are the main risk factors for upper gastrointestinal bleeding in these patients.⁵⁷ The most common source of bleeding is stress ulcers in the stomach. The ulcers develop from a combination of factors, including decreased blood flow to the mucosa causing ischemia and hypoperfusion-reperfusion injury.

Pharmacologic inhibition of gastric acid secretion has been shown to reduce upper gastrointestinal bleeding in acute liver failure. A histamine H₂ receptor blocker or proton pump inhibitor should be given to prevent gastrointestinal bleeding in patients with acute liver failure.^1,58

EXPERIMENTAL TREATMENTS

Artificial liver support systems

Membranes and dialysate solutions have been developed to remove toxic substances that are normally metabolized by the liver. Two of these—the molecular adsorbent recycling system (MARS) and the extracorporeal liver assist device (ELAD)—were developed in the late 1990s. MARS consisted of a highly permeable hollow fiber membrane mixed with albumin, and ELAD consisted of porcine hepatocytes attached to microcarriers in the extracapillary space of the hollow fiber membrane. Both systems allowed for transfer of water-soluble and protein-bound toxins in the blood across the membrane and into the dialysate.⁵⁹ The clinical benefit offered by these devices is controversial,^60–62 thus limiting their use to experimental purposes only.

Hepatocyte transplant

Use of hepatocyte transplant as a bridge to liver transplant was tested in 1970s, first in rats and later in humans.⁶³ By reducing the blood ammonia level and improving cerebral perfusion pressure and cardiac function, replacement of 1% to 2% of the total liver cell mass by transplanted hepatocytes acts as a bridge to orthotopic liver transplant.^64,65

PROGNOSIS

Different criteria have been used to identify patients with poor prognosis who may eventually need to undergo liver transplant.

The King’s College criteria system is the most commonly used for prognosis (Table 4).^37,66–69 Its main drawback is that it is applicable only in patients with encephalopathy, and when patients reach this stage, their condition often deteriorates rapidly, and they die while awaiting liver transplant.^37,66,67

The Model for End-Stage Liver Disease (MELD) score is an alternative to the King’s College criteria. A high MELD score on admission signifies advanced disease, and patients with a high MELD score tend to have a worse prognosis than those with a low score.⁶⁸

The Acute Physiology and Chronic Health Evaluation (APACHE) II score can also be used, as it is more sensitive than the King’s College criteria.⁶

The Clichy criteria^66,69 can also be used.

Liver biopsy. In addition to helping establish the cause of acute liver failure, liver biopsy can also be used as a prognostic tool. Hepatocellular necrosis greater than 70% on the biopsy predicts death with a specificity of 90% and a sensitivity of 56%.⁷⁰

Hypophosphatemia has been reported to indicate recovering liver function in patients with acute liver failure.⁷¹ As the liver regenerates, its energy requirement increases. To supply the energy, adenosine triphosphate production increases, and phosphorus shifts from the extracellular to the intracellular compartment to meet the need for extra phosphorus during this process. A serum phosphorus level of 2.9 mg/dL or higher appears to indicate a poor prognosis in patients with acute liver failure, as it signifies that adequate hepatocyte regeneration is not occurring.

Celiac disease is an autoimmune disorder that occurs in genetically predisposed individuals in response to ingestion of gluten. Its prevalence is about 0.7% of the US population.¹

See related editorial

The gold standard for diagnosis is duodenal biopsy, in which the histologic features may include varying gradations of flattening of intestinal villi, crypt hyperplasia, and infiltration of the lamina propria by lymphocytes. Many patients have no symptoms at the time of diagnosis, but presenting symptoms can include diarrhea along with features of malabsorption,² and, in about 25% of patients (mainly adults), a bullous cutaneous disorder called dermatitis herpetiformis.^3,4 The pathogenesis of celiac disease and that of dermatitis herpetiformis are similar in that in both, ingestion of gluten induces an inflammatory reaction leading to the clinical manifestations.

The mainstay of treatment of celiac disease remains avoidance of gluten in the diet.

GENETIC PREDISPOSITION AND DIETARY TRIGGER

The pathogenesis of celiac disease has been well studied in both humans and animals. The disease is thought to develop by an interplay of genetic and autoimmune factors and the ingestion of gluten (ie, an environmental factor).

Celiac disease occurs in genetically predisposed individuals, ie, those who carry the HLA alleles DQ2 (DQA1*05, DQB1*02), DQ8 (DQA1*03, DQB1*0302), or both.⁵

Ingestion of gluten is necessary for the disease to develop. Gluten, the protein component of wheat, barley, and rye, contains proteins called prolamins, which vary among the different types of grain. In wheat, the prolamin is gliadin, which is alcohol-soluble. In barley the prolamin is hordein, and in rye it is secalin.⁴ The prolamin content in gluten makes it resistant to degradation by gastric, pancreatic, and intestinal brush border proteases.⁶ Gluten crosses the epithelial barrier and promotes an inflammatory reaction by both the innate and adaptive immune systems that can ultimately result in flattening of villi and crypt hyperplasia (Figure 1).⁷

Tissue transglutaminase also plays a central role in the pathogenesis, as it further deaminates gliadin and increases its immunogenicity by causing it to bind to receptors on antigen-presenting cells with stronger affinity. Furthermore, gliadin-tissue transglutaminase complexes formed by protein cross-linkages generate an autoantibody response (predominantly immunoglobulin A [IgA] type) that can exacerbate the inflammatory process.^8,9

Certain viral infections during childhood, such as rotavirus and adenovirus infection, can increase the risk of celiac disease.^10–13 Although earlier studies reported that breast-feeding seemed to have a protective effect,¹⁴ as did introducing grains in the diet in the 4th to 6th months of life as opposed to earlier or later,¹⁵ more recent studies have not confirmed these benefits.^16,17

CLINICAL FEATURES

Most adults diagnosed with celiac disease are in their 30s, 40s, or 50s, and most are women.

Diarrhea remains a common presenting symptom, although the percentage of patients with celiac disease who present with diarrhea has decreased over time.^18,19

Abdominal pain and weight loss are also common.²⁰

Pallor or decreased exercise tolerance can develop due to anemia from iron malabsorption, and some patients have easy bruising due to vitamin K malabsorption.

Gynecologic and obstetric complications associated with celiac disease include delayed menarche, amenorrhea, spontaneous abortion, intrauterine growth retardation, preterm delivery, and low-birth-weight babies.^21,22 Patients who follow a gluten-free diet tend to have a lower incidence of intrauterine growth retardation, preterm delivery, and low-birth-weight babies compared with untreated patients.^21,22

Osteoporosis and osteopenia due to malabsorption of vitamin D are common and are seen in two-thirds of patients presenting with celiac disease.²³ A meta-analysis and position statement from Canada concluded that dual-energy x-ray absorptiometry should be done at the time of diagnosis of celiac disease if the patient is at risk of osteoporosis.²⁴ If the scan is abnormal, it should be repeated 1 to 2 years after initiation of a gluten-free diet and vitamin D supplementation to ensure that the osteopenia has improved.²⁴

OTHER DISEASE ASSOCIATIONS

Celiac disease is associated with various other autoimmune diseases (Table 1), including Hashimoto thyroiditis,²⁵ type 1 diabetes mellitus,²⁶ primary biliary cirrhosis,²⁷ primary sclerosing cholangitis,²⁸ and Addison disease.²⁹

Dermatitis herpetiformis

Dermatitis herpetiformis is one of the most common cutaneous manifestations of celiac disease. It presents between ages 10 and 50, and unlike celiac disease, it is more common in males.³⁰

Photo courtesy of Alok Vij, Department of Dermatology, Cleveland Clinic.

The characteristic lesions are pruritic, grouped erythematous papules surmounted by vesicles distributed symmetrically over the extensor surfaces of the upper and lower extremities, elbows, knees, scalp, nuchal area, and buttocks³¹ (Figures 2 and 3). In addition, some patients also present with vesicles, erythematous macules, and erosions in the oral mucosa³² or purpura on the palms and soles.^33–35

Photo courtesy of Alok Vij, MD, Department of Dermatology, Cleveland Clinic.

The pathogenesis of dermatitis herpetiformis in the skin is related to the pathogenesis of celiac disease in the gut. Like celiac disease, dermatitis herpetiformis is more common in genetically predisposed individuals carrying either the HLA-DQ2 or the HLA-DQ8 haplotype. In the skin, there is an analogue of tissue transglutaminase called epidermal transglutaminase, which helps in maintaining the integrity of cornified epithelium.³⁶ In patients with celiac disease, along with formation of IgA antibodies to tissue transglutaminase, there is also formation of IgA antibodies to epidermal transglutaminase. IgA antibodies are deposit- ed in the tips of dermal papillae and along the basement membrane.^37–39 These deposits then initiate an inflammatory response that is predominantly neutrophilic and results in formation of vesicles and bullae in the skin.⁴⁰ Also supporting the linkage between celiac disease and dermatitis herpetiformis, if patients adhere to a gluten-free diet, the deposits of immune complexes in the skin disappear.⁴¹

CELIAC DISEASE-ASSOCIATED MALIGNANCY

Patients with celiac disease have a higher risk of developing enteric malignancies, particularly intestinal T-cell lymphoma, and they have smaller increased risk of colon, oropharyngeal, esophageal, pancreatic, and hepatobiliary cancer.^42–45 For all of these cancers, the risk is higher than in the general public in the first year after celiac disease is diagnosed, but after the first year, the risk is increased only for small-bowel and hepatobiliary malignancies.⁴⁶

T-cell lymphoma

T-cell lymphoma is a rare but serious complication that has a poor prognosis.⁴⁷ Its prevalence has been increasing with time and is currently estimated to be around 0.01 to 0.02 per 100,000 people in the population as a whole.^48,49 The risk of developing lymphoma is 2.5 times higher in people with celiac disease than in the general population.⁵⁰ T-cell lymphoma is seen more commonly in patients with refractory celiac disease and DQ2 homozygosity.⁵¹

This disease is difficult to detect clinically, but sometimes it presents as an acute exacer­bation of celiac disease symptoms despite strict adherence to a gluten-free diet. Associated alarm symptoms include fever, night sweats, and laboratory abnormalities such as low albumin and high lactate dehydrogenase levels.

Strict adherence to a gluten-free diet remains the only way to prevent intestinal T-cell lymphoma.⁵²

Other malignancies

Some earlier studies reported an increased risk of thyroid cancer and malignant melanoma, but two newer studies have refuted this finding.^53,54 Conversely, celiac disease appears to have a protective effect against breast, ovarian, and endometrial cancers.⁵⁵

DIAGNOSIS: SEROLOGY, BIOPSY, GENETIC TESTING

Serologic tests

Patients strongly suspected of having celiac disease should be screened for IgA antibodies to tissue transglutaminase while on a gluten-containing diet, according to recommendations of the American College of Gastroenterology (Figure 4).⁵⁶ The sensitivity and specificity of this test are around 95%. If the patient has an IgA deficiency, screening should be done by checking the level of IgG antibodies to tissue transglutaminase.

Biopsy for confirmation

If testing for IgA to tissue transglutaminase is positive, upper endoscopy with biopsy is needed. Ideally, one to two samples should be taken from the duodenal bulb and at least four samples from the rest of the duodenum, preferably from two different locations.⁵⁶

Celiac disease has a broad spectrum of pathologic expressions, from mild distortion of crypt architecture to total villous atrophy and infiltration of lamina propria by lymphocytes⁵⁷ (Figures 5 and 6). Because these changes can be seen in a variety of diarrheal diseases, their reversal after adherence to a gluten-free diet is part of the current diagnostic criteria for the diagnosis of celiac disease.⁵⁶

Genetic testing

Photomicrograph courtesy of Homer Wiland MD, Department of Pathology, Cleveland Clinic.

Although the combination of positive serologic tests and pathologic changes confirms the diagnosis of celiac disease, in some cases one type of test is positive and the other is negative. In this situation, genetic testing for HLA-DQ2 and HLA-DQ8 can help rule out the diagnosis, as a negative genetic test rules out celiac disease in more than 99% of cases.⁵⁸

Genetic testing is also useful in patients who are already adhering to a gluten-free diet at the time of presentation to the clinic and who have had no testing done for celiac disease in the past. Here again, a negative test for both HLA-DQ2 and HLA-DQ8 makes a diagnosis of celiac disease highly unlikely.

If the test is positive, further testing needs to be done, as a positive genetic test cannot differentiate celiac disease from nonceliac gluten sensitivity. In this case, a gluten challenge needs to be done, ideally for 8 weeks, but for at least 2 weeks if the patient cannot tolerate gluten-containing food for a longer period of time. The gluten challenge is to be followed by testing for antibodies to tissue transglutaminase or obtaining duodenal biopsies to confirm the presence or absence of celiac disease.

Standard laboratory tests

Standard laboratory tests do not help much in diagnosing celiac disease, but they should include a complete blood chemistry along with a complete metabolic panel. Usually, serum albumin levels are normal.

Due to malabsorption of iron, patients may have iron deficiency anemia,⁵⁹ but anemia can also be due to a deficiency of folate or vitamin B₁₂. In patients undergoing endoscopic evaluation of iron deficiency anemia of unknown cause, celiac disease was discovered in approximately 15%.⁶⁰ Therefore, some experts believe that any patient presenting with unexplained iron deficiency anemia should be screened for celiac disease.

Because of malabsorption of vitamin D, levels of vitamin D can be low.

Elevations in levels of aminotransferases are also fairly common and usually resolve after the start of a gluten-free diet. If they persist despite adherence to a gluten-free diet, then an alternate cause of liver disease should be sought.⁶¹

Diagnosis of dermatitis herpetiformis

When trying to diagnose dermatitis herpetiformis, antibodies against epidermal transglutaminase can also be checked if testing for antibody against tissue transglutaminase is negative. A significant number of patients with biopsy-confirmed dermatitis herpetiformis are positive for epidermal transglutaminase antibodies but not for tissue transglutaminase antibodies.⁶²

The confirmatory test for dermatitis herpetiformis remains skin biopsy. Ideally, the sample should be taken while the patient is on a gluten-containing diet and from an area of normal-appearing skin around the lesions.⁶³ On histopathologic study, neutrophilic infiltrates are seen in dermal papillae and a perivascular lymphocytic infiltrate can also be seen in the superficial zones.⁶⁴ This presentation can also be seen in other bullous disorders, however. To differentiate dermatitis herpetiformis from other disorders, direct immunofluorescence is needed, which will detect granular IgA deposits in the dermal papillae or along the basement membrane, a finding pathognomic of dermatitis herpetiformis.⁶³

A GLUTEN-FREE DIET IS THE MAINSTAY OF TREATMENT

The mainstay of treatment is lifelong adherence to a gluten-free diet. Most patients report improvement in abdominal pain within days of starting this diet and improvement of diarrhea within 4 weeks.⁶⁵

The maximum amount of gluten that can be tolerated is debatable. A study established that intake of less than 10 mg a day is associated with fewer histologic abnormalities,⁶⁶ and an earlier study noted that intake of less than 50 mg a day was clinically well tolerated.⁶⁷ But patients differ in their tolerance for gluten, and it is hard to predict what the threshold of tolerance for gluten will be for a particular individual. Thus, it is better to avoid gluten completely.

Gluten-free if it is inherently gluten-free. If the food has a gluten-containing grain, then it should be processed to remove the gluten, and the resultant food product should not contain more than 20 parts per million of gluten. Gluten-free products that have gluten-containing grain that has been processed usually have a label indicating the gluten content in the food in parts per million.

Patients who understand the need to adhere to a gluten-free diet and the implications of not adhering to it are generally more compliant. Thus, patients need to be strongly educated that they need to adhere to a gluten-free diet and that nonadherence can cause further damage to the gut and can pose a higher risk of malignancy. Even though patients are usually concerned about the cost of gluten-free food and worry about adherence to the diet, these factors do not generally limit diet adherence.⁶⁸ All patients diagnosed with celiac disease should meet with a registered dietitian to discuss diet options based on their food preferences and to better address all their concerns.

With increasing awareness of celiac disease and with increasing numbers of patients being diagnosed with it, the food industry has recognized the need to produce gluten-free items. There are now plenty of food products available for these patients, who no longer have to forgo cakes, cookies, and other such items. Table 2 lists some common foods that patients with celiac disease can consume.

Nutritional supplements for some

If anemia is due purely to iron deficiency, it may resolve after starting a gluten-free diet, and no additional supplementation may be needed. However, if it is due to a combination of iron plus folate or vitamin B₁₂ deficiency, then folate, vitamin B₁₂, or both should be given.

In addition, if the patient is found to have a deficiency of vitamin D, then a vitamin D supplement should be given.⁶⁹ At the time of diagnosis, all patients with celiac disease should be screened for deficiencies of vitamins A, B₁₂, D, E, and K, as well as copper, zinc, folic acid, and iron.

Follow-up at 3 to 6 months

A follow-up visit should be scheduled for 3 to 6 months after the diagnosis and after that on an annual basis, and many of the abnormal laboratory tests will need to be repeated.

If intestinal or extraintestinal symptoms or nutrient deficiencies persist, then the patient’s adherence to the gluten-free diet needs to be checked. Adherence to a gluten-free diet can be assessed by checking for serologic markers of celiac disease. A decrease in baseline values can be seen within a few months of starting the diet.⁷⁰ Failure of serologic markers to decrease by the end of 1 year of a gluten-free diet usually indicates gluten contamination.⁷¹ If adherence is confirmed (ie, if baseline values fall) but symptoms persist, then further workup needs to be done to find the cause of refractory disease.

Skin lesions should also respond to a gluten-free diet

The first and foremost therapy for the skin lesions in dermatitis herpetiformis is the same as that for the intestinal manifestations in celiac disease, ie, adherence to a gluten-free diet. Soon after patients begin a gluten-free diet, the itching around the skin lesions goes away, and over time, most patients have complete resolution of the skin manifestations.

Dapsone is also frequently used to treat dermatitis herpetiformis if there is an incomplete response to a gluten-free diet or as an adjunct to diet to treat the pruritus. Patients often have a good response to dapsone.⁷²

The recommended starting dosage is 100 to 200 mg a day, and a response is usually seen within a few days. If the symptoms do not improve, the dose can be increased. Once the lesions resolve, the dose can be tapered and patients may not require any further medication. In some cases, patients may need to be chronically maintained on the lowest dose possible, due to the side effects of the drug.³

Dapsone is associated with significant adverse effects. Methemoglobinemia is the most common and is seen particularly in dosages exceeding 200 mg a day. Hemolytic anemia, another common adverse effect, is seen with dosages of more than 100 mg a day. Patients with a deficiency of glucose-6-phosphate dehydrogenase (G6PD) are at increased risk of hemolysis, and screening for G6PD deficiency is usually done before starting dapsone. Other rare adverse effects of dapsone include agranulocytosis, peripheral neuropathy, psychosis,⁷³ pancreatitis, cholestatic jaundice, bullous and exfoliative dermatitis, Stevens-Johnson syndrome, toxic epidermal necrolysis, nephrotic syndrome, and renal papillary necrosis.

Besides testing for G6PD deficiency, a complete blood cell count, a reticulocyte count, a hepatic function panel, renal function tests, and urinalysis should be done before starting dapsone therapy and repeated while on therapy. The complete blood cell count and reticulocyte count should be checked weekly for the first month, twice a month for the next 2 months, and then once every 3 months. Liver and renal function tests are to be done once every 3 months.⁷⁴

NOVEL THERAPIES BEING TESTED

Research is under way for other treatments for celiac disease besides a gluten-free diet.

Larazotide (Alba Therapeutics, Baltimore, MD) is being tested in a randomized, placebo-controlled trial. Early results indicate that it is effective in controlling both gastrointestinal and nongastrointestinal symptoms of celiac disease, but it still has to undergo phase 3 clinical trials.

Sorghum is a grain commonly used in Asia and Africa. The gluten in sorghum is different from that in wheat and is not immunogenic. In a small case series in patients with known celiac disease, sorghum did not induce diarrhea or change in levels of antibodies to tissue transglutaminase.⁷⁵

Nonimmunogenic wheat that does not contain the immunogenic gluten is being developed.

Oral enzyme supplements called glutenases are being developed. Glutenases can cleave gluten, particularly the proline and glutamine residues that make gluten resistant to degradation by gastric, pancreatic, and intestinal brush border proteases. A phase 2 trial of one of these oral enzyme supplements showed that it appeared to attenuate mucosal injury in patients with biopsy-proven celiac disease.⁷⁶

These novel therapies look promising, but for now the best treatment is lifelong adherence to the gluten-free diet.

NONRESPONSIVE AND REFRACTORY CELIAC DISEASE

Celiac disease is considered nonresponsive if its symptoms or laboratory abnormalities persist after the patient is on a gluten-free diet for 6 to 12 months. It is considered refractory if symptoms persist or recur along with villous atrophy despite adherence to the diet for more than 12 months in the absence of other causes of the symptoms. Refractory celiac disease can be further classified either as type 1 if there are typical intraepithelial lymphocytes, or as type 2 if there are atypical intraepithelial lymphocytes.

Celiac disease is nonresponsive in about 10% to 19% of cases,⁷⁶ and it is refractory in 1% to 2%.⁷⁷

Managing nonresponsive celiac disease

The first step in managing a patient with nonresponsive celiac disease is to confirm the diagnosis by reviewing the serologic tests and the biopsy samples from the time of diagnosis. If celiac disease is confirmed, then one should re-evaluate for gluten ingestion, the most common cause of nonresponsiveness.⁷⁸ If strict adherence is confirmed, then check for other causes of symptoms such as lactose or fructose intolerance. If no other cause is found, then repeat the duodenal biopsies with flow cytometry to look for CD3 and CD8 expression in T cells in the small-bowel mucosa.⁷⁹ Presence or absence of villous atrophy can point to possible other causes of malabsorption including pancreatic insufficiency, small intestinal bowel overgrowth, and microscopic colitis.

Managing refractory celiac disease

Traditionally, corticosteroids have been shown to be beneficial in alleviating symptoms in patients with refractory celiac disease but do not improve the histologic findings.⁸⁰ Because of the adverse effects associated with long-term corticosteroid use, azathioprine has been successfully used to maintain remission of the disease after induction with corticosteroids in patients with type 1 refractory celiac disease.⁸¹

Cladribine, a chemotherapeutic agent used to treat hairy cell leukemia, has shown some benefit in treating type 2 refractory celiac disease.⁸²

In type 2 refractory celiac disease, use of an immunomodulator agent carries an increased risk of transformation to lymphoma.

Because of the lack of a satisfactory response to the agents available so far to treat refractory celiac disease, more treatment options acting at the molecular level are being explored.

NONCELIAC GLUTEN SENSITIVITY DISORDER

Nonceliac gluten sensitivity disorder is an evolving concept. The clinical presentation of this disorder is similar to celiac disease in that patients may have diarrhea or other extra­intestinal symptoms when on a regular diet and have resolution of symptoms on a gluten-free diet. But unlike celiac disease, there is no serologic or histologic evidence of celiac disease even when patients are on a regular diet.

One of every 17 patients who presents with clinical features suggestive of celiac disease is found to have nonceliac gluten sensitivity disorder, not celiac disease.⁸³ In contrast to celiac disease, in which the adaptive immune system is thought to contribute to the disease process, in nonceliac gluten sensitivity disorder the innate immune system is believed to play the dominant role,⁸⁴ but the exact pathogenesis of the disease is still unclear.

The diagnosis of nonceliac gluten sensitivity disorder is one of exclusion. Celiac disease needs to be ruled out by serologic testing and by duodenal biopsy while the patient is on a regular diet, and then a trial of a gluten-free diet needs to be done to confirm resolution of symptoms before the diagnosis of nonceliac gluten sensitivity disorder can be established.

As with celiac disease, the treatment involves adhering to a gluten-free diet, but it is still not known if patients need to stay on it for the rest of their life, or if they will be able to tolerate gluten-containing products after a few years.

Celiac disease is an autoimmune disorder that occurs in genetically predisposed individuals in response to ingestion of gluten. Its prevalence is about 0.7% of the US population.¹

See related editorial

The gold standard for diagnosis is duodenal biopsy, in which the histologic features may include varying gradations of flattening of intestinal villi, crypt hyperplasia, and infiltration of the lamina propria by lymphocytes. Many patients have no symptoms at the time of diagnosis, but presenting symptoms can include diarrhea along with features of malabsorption,² and, in about 25% of patients (mainly adults), a bullous cutaneous disorder called dermatitis herpetiformis.^3,4 The pathogenesis of celiac disease and that of dermatitis herpetiformis are similar in that in both, ingestion of gluten induces an inflammatory reaction leading to the clinical manifestations.

The mainstay of treatment of celiac disease remains avoidance of gluten in the diet.

GENETIC PREDISPOSITION AND DIETARY TRIGGER

The pathogenesis of celiac disease has been well studied in both humans and animals. The disease is thought to develop by an interplay of genetic and autoimmune factors and the ingestion of gluten (ie, an environmental factor).

Celiac disease occurs in genetically predisposed individuals, ie, those who carry the HLA alleles DQ2 (DQA1*05, DQB1*02), DQ8 (DQA1*03, DQB1*0302), or both.⁵

Ingestion of gluten is necessary for the disease to develop. Gluten, the protein component of wheat, barley, and rye, contains proteins called prolamins, which vary among the different types of grain. In wheat, the prolamin is gliadin, which is alcohol-soluble. In barley the prolamin is hordein, and in rye it is secalin.⁴ The prolamin content in gluten makes it resistant to degradation by gastric, pancreatic, and intestinal brush border proteases.⁶ Gluten crosses the epithelial barrier and promotes an inflammatory reaction by both the innate and adaptive immune systems that can ultimately result in flattening of villi and crypt hyperplasia (Figure 1).⁷

Tissue transglutaminase also plays a central role in the pathogenesis, as it further deaminates gliadin and increases its immunogenicity by causing it to bind to receptors on antigen-presenting cells with stronger affinity. Furthermore, gliadin-tissue transglutaminase complexes formed by protein cross-linkages generate an autoantibody response (predominantly immunoglobulin A [IgA] type) that can exacerbate the inflammatory process.^8,9

Certain viral infections during childhood, such as rotavirus and adenovirus infection, can increase the risk of celiac disease.^10–13 Although earlier studies reported that breast-feeding seemed to have a protective effect,¹⁴ as did introducing grains in the diet in the 4th to 6th months of life as opposed to earlier or later,¹⁵ more recent studies have not confirmed these benefits.^16,17

CLINICAL FEATURES

Most adults diagnosed with celiac disease are in their 30s, 40s, or 50s, and most are women.

Diarrhea remains a common presenting symptom, although the percentage of patients with celiac disease who present with diarrhea has decreased over time.^18,19

Abdominal pain and weight loss are also common.²⁰

Pallor or decreased exercise tolerance can develop due to anemia from iron malabsorption, and some patients have easy bruising due to vitamin K malabsorption.

Gynecologic and obstetric complications associated with celiac disease include delayed menarche, amenorrhea, spontaneous abortion, intrauterine growth retardation, preterm delivery, and low-birth-weight babies.^21,22 Patients who follow a gluten-free diet tend to have a lower incidence of intrauterine growth retardation, preterm delivery, and low-birth-weight babies compared with untreated patients.^21,22

Osteoporosis and osteopenia due to malabsorption of vitamin D are common and are seen in two-thirds of patients presenting with celiac disease.²³ A meta-analysis and position statement from Canada concluded that dual-energy x-ray absorptiometry should be done at the time of diagnosis of celiac disease if the patient is at risk of osteoporosis.²⁴ If the scan is abnormal, it should be repeated 1 to 2 years after initiation of a gluten-free diet and vitamin D supplementation to ensure that the osteopenia has improved.²⁴

OTHER DISEASE ASSOCIATIONS

Celiac disease is associated with various other autoimmune diseases (Table 1), including Hashimoto thyroiditis,²⁵ type 1 diabetes mellitus,²⁶ primary biliary cirrhosis,²⁷ primary sclerosing cholangitis,²⁸ and Addison disease.²⁹

Dermatitis herpetiformis

Dermatitis herpetiformis is one of the most common cutaneous manifestations of celiac disease. It presents between ages 10 and 50, and unlike celiac disease, it is more common in males.³⁰

Photo courtesy of Alok Vij, Department of Dermatology, Cleveland Clinic.

The characteristic lesions are pruritic, grouped erythematous papules surmounted by vesicles distributed symmetrically over the extensor surfaces of the upper and lower extremities, elbows, knees, scalp, nuchal area, and buttocks³¹ (Figures 2 and 3). In addition, some patients also present with vesicles, erythematous macules, and erosions in the oral mucosa³² or purpura on the palms and soles.^33–35

Photo courtesy of Alok Vij, MD, Department of Dermatology, Cleveland Clinic.

The pathogenesis of dermatitis herpetiformis in the skin is related to the pathogenesis of celiac disease in the gut. Like celiac disease, dermatitis herpetiformis is more common in genetically predisposed individuals carrying either the HLA-DQ2 or the HLA-DQ8 haplotype. In the skin, there is an analogue of tissue transglutaminase called epidermal transglutaminase, which helps in maintaining the integrity of cornified epithelium.³⁶ In patients with celiac disease, along with formation of IgA antibodies to tissue transglutaminase, there is also formation of IgA antibodies to epidermal transglutaminase. IgA antibodies are deposit- ed in the tips of dermal papillae and along the basement membrane.^37–39 These deposits then initiate an inflammatory response that is predominantly neutrophilic and results in formation of vesicles and bullae in the skin.⁴⁰ Also supporting the linkage between celiac disease and dermatitis herpetiformis, if patients adhere to a gluten-free diet, the deposits of immune complexes in the skin disappear.⁴¹

CELIAC DISEASE-ASSOCIATED MALIGNANCY

Patients with celiac disease have a higher risk of developing enteric malignancies, particularly intestinal T-cell lymphoma, and they have smaller increased risk of colon, oropharyngeal, esophageal, pancreatic, and hepatobiliary cancer.^42–45 For all of these cancers, the risk is higher than in the general public in the first year after celiac disease is diagnosed, but after the first year, the risk is increased only for small-bowel and hepatobiliary malignancies.⁴⁶

T-cell lymphoma

T-cell lymphoma is a rare but serious complication that has a poor prognosis.⁴⁷ Its prevalence has been increasing with time and is currently estimated to be around 0.01 to 0.02 per 100,000 people in the population as a whole.^48,49 The risk of developing lymphoma is 2.5 times higher in people with celiac disease than in the general population.⁵⁰ T-cell lymphoma is seen more commonly in patients with refractory celiac disease and DQ2 homozygosity.⁵¹

This disease is difficult to detect clinically, but sometimes it presents as an acute exacer­bation of celiac disease symptoms despite strict adherence to a gluten-free diet. Associated alarm symptoms include fever, night sweats, and laboratory abnormalities such as low albumin and high lactate dehydrogenase levels.

Strict adherence to a gluten-free diet remains the only way to prevent intestinal T-cell lymphoma.⁵²

Other malignancies

Some earlier studies reported an increased risk of thyroid cancer and malignant melanoma, but two newer studies have refuted this finding.^53,54 Conversely, celiac disease appears to have a protective effect against breast, ovarian, and endometrial cancers.⁵⁵

DIAGNOSIS: SEROLOGY, BIOPSY, GENETIC TESTING

Serologic tests

Patients strongly suspected of having celiac disease should be screened for IgA antibodies to tissue transglutaminase while on a gluten-containing diet, according to recommendations of the American College of Gastroenterology (Figure 4).⁵⁶ The sensitivity and specificity of this test are around 95%. If the patient has an IgA deficiency, screening should be done by checking the level of IgG antibodies to tissue transglutaminase.

Biopsy for confirmation

If testing for IgA to tissue transglutaminase is positive, upper endoscopy with biopsy is needed. Ideally, one to two samples should be taken from the duodenal bulb and at least four samples from the rest of the duodenum, preferably from two different locations.⁵⁶

Celiac disease has a broad spectrum of pathologic expressions, from mild distortion of crypt architecture to total villous atrophy and infiltration of lamina propria by lymphocytes⁵⁷ (Figures 5 and 6). Because these changes can be seen in a variety of diarrheal diseases, their reversal after adherence to a gluten-free diet is part of the current diagnostic criteria for the diagnosis of celiac disease.⁵⁶

Genetic testing

Photomicrograph courtesy of Homer Wiland MD, Department of Pathology, Cleveland Clinic.

Although the combination of positive serologic tests and pathologic changes confirms the diagnosis of celiac disease, in some cases one type of test is positive and the other is negative. In this situation, genetic testing for HLA-DQ2 and HLA-DQ8 can help rule out the diagnosis, as a negative genetic test rules out celiac disease in more than 99% of cases.⁵⁸

Genetic testing is also useful in patients who are already adhering to a gluten-free diet at the time of presentation to the clinic and who have had no testing done for celiac disease in the past. Here again, a negative test for both HLA-DQ2 and HLA-DQ8 makes a diagnosis of celiac disease highly unlikely.

If the test is positive, further testing needs to be done, as a positive genetic test cannot differentiate celiac disease from nonceliac gluten sensitivity. In this case, a gluten challenge needs to be done, ideally for 8 weeks, but for at least 2 weeks if the patient cannot tolerate gluten-containing food for a longer period of time. The gluten challenge is to be followed by testing for antibodies to tissue transglutaminase or obtaining duodenal biopsies to confirm the presence or absence of celiac disease.

Standard laboratory tests

Standard laboratory tests do not help much in diagnosing celiac disease, but they should include a complete blood chemistry along with a complete metabolic panel. Usually, serum albumin levels are normal.

Due to malabsorption of iron, patients may have iron deficiency anemia,⁵⁹ but anemia can also be due to a deficiency of folate or vitamin B₁₂. In patients undergoing endoscopic evaluation of iron deficiency anemia of unknown cause, celiac disease was discovered in approximately 15%.⁶⁰ Therefore, some experts believe that any patient presenting with unexplained iron deficiency anemia should be screened for celiac disease.

Because of malabsorption of vitamin D, levels of vitamin D can be low.

Elevations in levels of aminotransferases are also fairly common and usually resolve after the start of a gluten-free diet. If they persist despite adherence to a gluten-free diet, then an alternate cause of liver disease should be sought.⁶¹

Diagnosis of dermatitis herpetiformis

When trying to diagnose dermatitis herpetiformis, antibodies against epidermal transglutaminase can also be checked if testing for antibody against tissue transglutaminase is negative. A significant number of patients with biopsy-confirmed dermatitis herpetiformis are positive for epidermal transglutaminase antibodies but not for tissue transglutaminase antibodies.⁶²

The confirmatory test for dermatitis herpetiformis remains skin biopsy. Ideally, the sample should be taken while the patient is on a gluten-containing diet and from an area of normal-appearing skin around the lesions.⁶³ On histopathologic study, neutrophilic infiltrates are seen in dermal papillae and a perivascular lymphocytic infiltrate can also be seen in the superficial zones.⁶⁴ This presentation can also be seen in other bullous disorders, however. To differentiate dermatitis herpetiformis from other disorders, direct immunofluorescence is needed, which will detect granular IgA deposits in the dermal papillae or along the basement membrane, a finding pathognomic of dermatitis herpetiformis.⁶³

A GLUTEN-FREE DIET IS THE MAINSTAY OF TREATMENT

The mainstay of treatment is lifelong adherence to a gluten-free diet. Most patients report improvement in abdominal pain within days of starting this diet and improvement of diarrhea within 4 weeks.⁶⁵

The maximum amount of gluten that can be tolerated is debatable. A study established that intake of less than 10 mg a day is associated with fewer histologic abnormalities,⁶⁶ and an earlier study noted that intake of less than 50 mg a day was clinically well tolerated.⁶⁷ But patients differ in their tolerance for gluten, and it is hard to predict what the threshold of tolerance for gluten will be for a particular individual. Thus, it is better to avoid gluten completely.

Gluten-free if it is inherently gluten-free. If the food has a gluten-containing grain, then it should be processed to remove the gluten, and the resultant food product should not contain more than 20 parts per million of gluten. Gluten-free products that have gluten-containing grain that has been processed usually have a label indicating the gluten content in the food in parts per million.

Patients who understand the need to adhere to a gluten-free diet and the implications of not adhering to it are generally more compliant. Thus, patients need to be strongly educated that they need to adhere to a gluten-free diet and that nonadherence can cause further damage to the gut and can pose a higher risk of malignancy. Even though patients are usually concerned about the cost of gluten-free food and worry about adherence to the diet, these factors do not generally limit diet adherence.⁶⁸ All patients diagnosed with celiac disease should meet with a registered dietitian to discuss diet options based on their food preferences and to better address all their concerns.

With increasing awareness of celiac disease and with increasing numbers of patients being diagnosed with it, the food industry has recognized the need to produce gluten-free items. There are now plenty of food products available for these patients, who no longer have to forgo cakes, cookies, and other such items. Table 2 lists some common foods that patients with celiac disease can consume.

Nutritional supplements for some

If anemia is due purely to iron deficiency, it may resolve after starting a gluten-free diet, and no additional supplementation may be needed. However, if it is due to a combination of iron plus folate or vitamin B₁₂ deficiency, then folate, vitamin B₁₂, or both should be given.

In addition, if the patient is found to have a deficiency of vitamin D, then a vitamin D supplement should be given.⁶⁹ At the time of diagnosis, all patients with celiac disease should be screened for deficiencies of vitamins A, B₁₂, D, E, and K, as well as copper, zinc, folic acid, and iron.

Follow-up at 3 to 6 months

A follow-up visit should be scheduled for 3 to 6 months after the diagnosis and after that on an annual basis, and many of the abnormal laboratory tests will need to be repeated.

If intestinal or extraintestinal symptoms or nutrient deficiencies persist, then the patient’s adherence to the gluten-free diet needs to be checked. Adherence to a gluten-free diet can be assessed by checking for serologic markers of celiac disease. A decrease in baseline values can be seen within a few months of starting the diet.⁷⁰ Failure of serologic markers to decrease by the end of 1 year of a gluten-free diet usually indicates gluten contamination.⁷¹ If adherence is confirmed (ie, if baseline values fall) but symptoms persist, then further workup needs to be done to find the cause of refractory disease.

Skin lesions should also respond to a gluten-free diet

The first and foremost therapy for the skin lesions in dermatitis herpetiformis is the same as that for the intestinal manifestations in celiac disease, ie, adherence to a gluten-free diet. Soon after patients begin a gluten-free diet, the itching around the skin lesions goes away, and over time, most patients have complete resolution of the skin manifestations.

Dapsone is also frequently used to treat dermatitis herpetiformis if there is an incomplete response to a gluten-free diet or as an adjunct to diet to treat the pruritus. Patients often have a good response to dapsone.⁷²

The recommended starting dosage is 100 to 200 mg a day, and a response is usually seen within a few days. If the symptoms do not improve, the dose can be increased. Once the lesions resolve, the dose can be tapered and patients may not require any further medication. In some cases, patients may need to be chronically maintained on the lowest dose possible, due to the side effects of the drug.³

Dapsone is associated with significant adverse effects. Methemoglobinemia is the most common and is seen particularly in dosages exceeding 200 mg a day. Hemolytic anemia, another common adverse effect, is seen with dosages of more than 100 mg a day. Patients with a deficiency of glucose-6-phosphate dehydrogenase (G6PD) are at increased risk of hemolysis, and screening for G6PD deficiency is usually done before starting dapsone. Other rare adverse effects of dapsone include agranulocytosis, peripheral neuropathy, psychosis,⁷³ pancreatitis, cholestatic jaundice, bullous and exfoliative dermatitis, Stevens-Johnson syndrome, toxic epidermal necrolysis, nephrotic syndrome, and renal papillary necrosis.

Besides testing for G6PD deficiency, a complete blood cell count, a reticulocyte count, a hepatic function panel, renal function tests, and urinalysis should be done before starting dapsone therapy and repeated while on therapy. The complete blood cell count and reticulocyte count should be checked weekly for the first month, twice a month for the next 2 months, and then once every 3 months. Liver and renal function tests are to be done once every 3 months.⁷⁴

NOVEL THERAPIES BEING TESTED

Research is under way for other treatments for celiac disease besides a gluten-free diet.

Larazotide (Alba Therapeutics, Baltimore, MD) is being tested in a randomized, placebo-controlled trial. Early results indicate that it is effective in controlling both gastrointestinal and nongastrointestinal symptoms of celiac disease, but it still has to undergo phase 3 clinical trials.

Sorghum is a grain commonly used in Asia and Africa. The gluten in sorghum is different from that in wheat and is not immunogenic. In a small case series in patients with known celiac disease, sorghum did not induce diarrhea or change in levels of antibodies to tissue transglutaminase.⁷⁵

Nonimmunogenic wheat that does not contain the immunogenic gluten is being developed.

Oral enzyme supplements called glutenases are being developed. Glutenases can cleave gluten, particularly the proline and glutamine residues that make gluten resistant to degradation by gastric, pancreatic, and intestinal brush border proteases. A phase 2 trial of one of these oral enzyme supplements showed that it appeared to attenuate mucosal injury in patients with biopsy-proven celiac disease.⁷⁶

These novel therapies look promising, but for now the best treatment is lifelong adherence to the gluten-free diet.

NONRESPONSIVE AND REFRACTORY CELIAC DISEASE

Celiac disease is considered nonresponsive if its symptoms or laboratory abnormalities persist after the patient is on a gluten-free diet for 6 to 12 months. It is considered refractory if symptoms persist or recur along with villous atrophy despite adherence to the diet for more than 12 months in the absence of other causes of the symptoms. Refractory celiac disease can be further classified either as type 1 if there are typical intraepithelial lymphocytes, or as type 2 if there are atypical intraepithelial lymphocytes.

Celiac disease is nonresponsive in about 10% to 19% of cases,⁷⁶ and it is refractory in 1% to 2%.⁷⁷

Managing nonresponsive celiac disease

The first step in managing a patient with nonresponsive celiac disease is to confirm the diagnosis by reviewing the serologic tests and the biopsy samples from the time of diagnosis. If celiac disease is confirmed, then one should re-evaluate for gluten ingestion, the most common cause of nonresponsiveness.⁷⁸ If strict adherence is confirmed, then check for other causes of symptoms such as lactose or fructose intolerance. If no other cause is found, then repeat the duodenal biopsies with flow cytometry to look for CD3 and CD8 expression in T cells in the small-bowel mucosa.⁷⁹ Presence or absence of villous atrophy can point to possible other causes of malabsorption including pancreatic insufficiency, small intestinal bowel overgrowth, and microscopic colitis.

Managing refractory celiac disease

Traditionally, corticosteroids have been shown to be beneficial in alleviating symptoms in patients with refractory celiac disease but do not improve the histologic findings.⁸⁰ Because of the adverse effects associated with long-term corticosteroid use, azathioprine has been successfully used to maintain remission of the disease after induction with corticosteroids in patients with type 1 refractory celiac disease.⁸¹

Cladribine, a chemotherapeutic agent used to treat hairy cell leukemia, has shown some benefit in treating type 2 refractory celiac disease.⁸²

In type 2 refractory celiac disease, use of an immunomodulator agent carries an increased risk of transformation to lymphoma.

Because of the lack of a satisfactory response to the agents available so far to treat refractory celiac disease, more treatment options acting at the molecular level are being explored.

NONCELIAC GLUTEN SENSITIVITY DISORDER

Nonceliac gluten sensitivity disorder is an evolving concept. The clinical presentation of this disorder is similar to celiac disease in that patients may have diarrhea or other extra­intestinal symptoms when on a regular diet and have resolution of symptoms on a gluten-free diet. But unlike celiac disease, there is no serologic or histologic evidence of celiac disease even when patients are on a regular diet.

One of every 17 patients who presents with clinical features suggestive of celiac disease is found to have nonceliac gluten sensitivity disorder, not celiac disease.⁸³ In contrast to celiac disease, in which the adaptive immune system is thought to contribute to the disease process, in nonceliac gluten sensitivity disorder the innate immune system is believed to play the dominant role,⁸⁴ but the exact pathogenesis of the disease is still unclear.

The diagnosis of nonceliac gluten sensitivity disorder is one of exclusion. Celiac disease needs to be ruled out by serologic testing and by duodenal biopsy while the patient is on a regular diet, and then a trial of a gluten-free diet needs to be done to confirm resolution of symptoms before the diagnosis of nonceliac gluten sensitivity disorder can be established.

As with celiac disease, the treatment involves adhering to a gluten-free diet, but it is still not known if patients need to stay on it for the rest of their life, or if they will be able to tolerate gluten-containing products after a few years.

Celiac disease is an autoimmune disorder that occurs in genetically predisposed individuals in response to ingestion of gluten. Its prevalence is about 0.7% of the US population.¹

See related editorial

The gold standard for diagnosis is duodenal biopsy, in which the histologic features may include varying gradations of flattening of intestinal villi, crypt hyperplasia, and infiltration of the lamina propria by lymphocytes. Many patients have no symptoms at the time of diagnosis, but presenting symptoms can include diarrhea along with features of malabsorption,² and, in about 25% of patients (mainly adults), a bullous cutaneous disorder called dermatitis herpetiformis.^3,4 The pathogenesis of celiac disease and that of dermatitis herpetiformis are similar in that in both, ingestion of gluten induces an inflammatory reaction leading to the clinical manifestations.

The mainstay of treatment of celiac disease remains avoidance of gluten in the diet.

GENETIC PREDISPOSITION AND DIETARY TRIGGER

The pathogenesis of celiac disease has been well studied in both humans and animals. The disease is thought to develop by an interplay of genetic and autoimmune factors and the ingestion of gluten (ie, an environmental factor).

Celiac disease occurs in genetically predisposed individuals, ie, those who carry the HLA alleles DQ2 (DQA1*05, DQB1*02), DQ8 (DQA1*03, DQB1*0302), or both.⁵

Ingestion of gluten is necessary for the disease to develop. Gluten, the protein component of wheat, barley, and rye, contains proteins called prolamins, which vary among the different types of grain. In wheat, the prolamin is gliadin, which is alcohol-soluble. In barley the prolamin is hordein, and in rye it is secalin.⁴ The prolamin content in gluten makes it resistant to degradation by gastric, pancreatic, and intestinal brush border proteases.⁶ Gluten crosses the epithelial barrier and promotes an inflammatory reaction by both the innate and adaptive immune systems that can ultimately result in flattening of villi and crypt hyperplasia (Figure 1).⁷

Tissue transglutaminase also plays a central role in the pathogenesis, as it further deaminates gliadin and increases its immunogenicity by causing it to bind to receptors on antigen-presenting cells with stronger affinity. Furthermore, gliadin-tissue transglutaminase complexes formed by protein cross-linkages generate an autoantibody response (predominantly immunoglobulin A [IgA] type) that can exacerbate the inflammatory process.^8,9

Certain viral infections during childhood, such as rotavirus and adenovirus infection, can increase the risk of celiac disease.^10–13 Although earlier studies reported that breast-feeding seemed to have a protective effect,¹⁴ as did introducing grains in the diet in the 4th to 6th months of life as opposed to earlier or later,¹⁵ more recent studies have not confirmed these benefits.^16,17

CLINICAL FEATURES

Most adults diagnosed with celiac disease are in their 30s, 40s, or 50s, and most are women.

Diarrhea remains a common presenting symptom, although the percentage of patients with celiac disease who present with diarrhea has decreased over time.^18,19

Abdominal pain and weight loss are also common.²⁰

Pallor or decreased exercise tolerance can develop due to anemia from iron malabsorption, and some patients have easy bruising due to vitamin K malabsorption.

Gynecologic and obstetric complications associated with celiac disease include delayed menarche, amenorrhea, spontaneous abortion, intrauterine growth retardation, preterm delivery, and low-birth-weight babies.^21,22 Patients who follow a gluten-free diet tend to have a lower incidence of intrauterine growth retardation, preterm delivery, and low-birth-weight babies compared with untreated patients.^21,22

Osteoporosis and osteopenia due to malabsorption of vitamin D are common and are seen in two-thirds of patients presenting with celiac disease.²³ A meta-analysis and position statement from Canada concluded that dual-energy x-ray absorptiometry should be done at the time of diagnosis of celiac disease if the patient is at risk of osteoporosis.²⁴ If the scan is abnormal, it should be repeated 1 to 2 years after initiation of a gluten-free diet and vitamin D supplementation to ensure that the osteopenia has improved.²⁴

OTHER DISEASE ASSOCIATIONS

Celiac disease is associated with various other autoimmune diseases (Table 1), including Hashimoto thyroiditis,²⁵ type 1 diabetes mellitus,²⁶ primary biliary cirrhosis,²⁷ primary sclerosing cholangitis,²⁸ and Addison disease.²⁹

Dermatitis herpetiformis

Dermatitis herpetiformis is one of the most common cutaneous manifestations of celiac disease. It presents between ages 10 and 50, and unlike celiac disease, it is more common in males.³⁰

Photo courtesy of Alok Vij, Department of Dermatology, Cleveland Clinic.

The characteristic lesions are pruritic, grouped erythematous papules surmounted by vesicles distributed symmetrically over the extensor surfaces of the upper and lower extremities, elbows, knees, scalp, nuchal area, and buttocks³¹ (Figures 2 and 3). In addition, some patients also present with vesicles, erythematous macules, and erosions in the oral mucosa³² or purpura on the palms and soles.^33–35

Photo courtesy of Alok Vij, MD, Department of Dermatology, Cleveland Clinic.

The pathogenesis of dermatitis herpetiformis in the skin is related to the pathogenesis of celiac disease in the gut. Like celiac disease, dermatitis herpetiformis is more common in genetically predisposed individuals carrying either the HLA-DQ2 or the HLA-DQ8 haplotype. In the skin, there is an analogue of tissue transglutaminase called epidermal transglutaminase, which helps in maintaining the integrity of cornified epithelium.³⁶ In patients with celiac disease, along with formation of IgA antibodies to tissue transglutaminase, there is also formation of IgA antibodies to epidermal transglutaminase. IgA antibodies are deposit- ed in the tips of dermal papillae and along the basement membrane.^37–39 These deposits then initiate an inflammatory response that is predominantly neutrophilic and results in formation of vesicles and bullae in the skin.⁴⁰ Also supporting the linkage between celiac disease and dermatitis herpetiformis, if patients adhere to a gluten-free diet, the deposits of immune complexes in the skin disappear.⁴¹

CELIAC DISEASE-ASSOCIATED MALIGNANCY

Patients with celiac disease have a higher risk of developing enteric malignancies, particularly intestinal T-cell lymphoma, and they have smaller increased risk of colon, oropharyngeal, esophageal, pancreatic, and hepatobiliary cancer.^42–45 For all of these cancers, the risk is higher than in the general public in the first year after celiac disease is diagnosed, but after the first year, the risk is increased only for small-bowel and hepatobiliary malignancies.⁴⁶

T-cell lymphoma

T-cell lymphoma is a rare but serious complication that has a poor prognosis.⁴⁷ Its prevalence has been increasing with time and is currently estimated to be around 0.01 to 0.02 per 100,000 people in the population as a whole.^48,49 The risk of developing lymphoma is 2.5 times higher in people with celiac disease than in the general population.⁵⁰ T-cell lymphoma is seen more commonly in patients with refractory celiac disease and DQ2 homozygosity.⁵¹

This disease is difficult to detect clinically, but sometimes it presents as an acute exacer­bation of celiac disease symptoms despite strict adherence to a gluten-free diet. Associated alarm symptoms include fever, night sweats, and laboratory abnormalities such as low albumin and high lactate dehydrogenase levels.

Strict adherence to a gluten-free diet remains the only way to prevent intestinal T-cell lymphoma.⁵²

Other malignancies

Some earlier studies reported an increased risk of thyroid cancer and malignant melanoma, but two newer studies have refuted this finding.^53,54 Conversely, celiac disease appears to have a protective effect against breast, ovarian, and endometrial cancers.⁵⁵

DIAGNOSIS: SEROLOGY, BIOPSY, GENETIC TESTING

Serologic tests

Patients strongly suspected of having celiac disease should be screened for IgA antibodies to tissue transglutaminase while on a gluten-containing diet, according to recommendations of the American College of Gastroenterology (Figure 4).⁵⁶ The sensitivity and specificity of this test are around 95%. If the patient has an IgA deficiency, screening should be done by checking the level of IgG antibodies to tissue transglutaminase.

Biopsy for confirmation

If testing for IgA to tissue transglutaminase is positive, upper endoscopy with biopsy is needed. Ideally, one to two samples should be taken from the duodenal bulb and at least four samples from the rest of the duodenum, preferably from two different locations.⁵⁶

Celiac disease has a broad spectrum of pathologic expressions, from mild distortion of crypt architecture to total villous atrophy and infiltration of lamina propria by lymphocytes⁵⁷ (Figures 5 and 6). Because these changes can be seen in a variety of diarrheal diseases, their reversal after adherence to a gluten-free diet is part of the current diagnostic criteria for the diagnosis of celiac disease.⁵⁶

Genetic testing

Photomicrograph courtesy of Homer Wiland MD, Department of Pathology, Cleveland Clinic.

Although the combination of positive serologic tests and pathologic changes confirms the diagnosis of celiac disease, in some cases one type of test is positive and the other is negative. In this situation, genetic testing for HLA-DQ2 and HLA-DQ8 can help rule out the diagnosis, as a negative genetic test rules out celiac disease in more than 99% of cases.⁵⁸

Genetic testing is also useful in patients who are already adhering to a gluten-free diet at the time of presentation to the clinic and who have had no testing done for celiac disease in the past. Here again, a negative test for both HLA-DQ2 and HLA-DQ8 makes a diagnosis of celiac disease highly unlikely.

If the test is positive, further testing needs to be done, as a positive genetic test cannot differentiate celiac disease from nonceliac gluten sensitivity. In this case, a gluten challenge needs to be done, ideally for 8 weeks, but for at least 2 weeks if the patient cannot tolerate gluten-containing food for a longer period of time. The gluten challenge is to be followed by testing for antibodies to tissue transglutaminase or obtaining duodenal biopsies to confirm the presence or absence of celiac disease.

Standard laboratory tests

Standard laboratory tests do not help much in diagnosing celiac disease, but they should include a complete blood chemistry along with a complete metabolic panel. Usually, serum albumin levels are normal.

Due to malabsorption of iron, patients may have iron deficiency anemia,⁵⁹ but anemia can also be due to a deficiency of folate or vitamin B₁₂. In patients undergoing endoscopic evaluation of iron deficiency anemia of unknown cause, celiac disease was discovered in approximately 15%.⁶⁰ Therefore, some experts believe that any patient presenting with unexplained iron deficiency anemia should be screened for celiac disease.

Because of malabsorption of vitamin D, levels of vitamin D can be low.

Elevations in levels of aminotransferases are also fairly common and usually resolve after the start of a gluten-free diet. If they persist despite adherence to a gluten-free diet, then an alternate cause of liver disease should be sought.⁶¹

Diagnosis of dermatitis herpetiformis

When trying to diagnose dermatitis herpetiformis, antibodies against epidermal transglutaminase can also be checked if testing for antibody against tissue transglutaminase is negative. A significant number of patients with biopsy-confirmed dermatitis herpetiformis are positive for epidermal transglutaminase antibodies but not for tissue transglutaminase antibodies.⁶²

The confirmatory test for dermatitis herpetiformis remains skin biopsy. Ideally, the sample should be taken while the patient is on a gluten-containing diet and from an area of normal-appearing skin around the lesions.⁶³ On histopathologic study, neutrophilic infiltrates are seen in dermal papillae and a perivascular lymphocytic infiltrate can also be seen in the superficial zones.⁶⁴ This presentation can also be seen in other bullous disorders, however. To differentiate dermatitis herpetiformis from other disorders, direct immunofluorescence is needed, which will detect granular IgA deposits in the dermal papillae or along the basement membrane, a finding pathognomic of dermatitis herpetiformis.⁶³

A GLUTEN-FREE DIET IS THE MAINSTAY OF TREATMENT

The mainstay of treatment is lifelong adherence to a gluten-free diet. Most patients report improvement in abdominal pain within days of starting this diet and improvement of diarrhea within 4 weeks.⁶⁵

The maximum amount of gluten that can be tolerated is debatable. A study established that intake of less than 10 mg a day is associated with fewer histologic abnormalities,⁶⁶ and an earlier study noted that intake of less than 50 mg a day was clinically well tolerated.⁶⁷ But patients differ in their tolerance for gluten, and it is hard to predict what the threshold of tolerance for gluten will be for a particular individual. Thus, it is better to avoid gluten completely.

Gluten-free if it is inherently gluten-free. If the food has a gluten-containing grain, then it should be processed to remove the gluten, and the resultant food product should not contain more than 20 parts per million of gluten. Gluten-free products that have gluten-containing grain that has been processed usually have a label indicating the gluten content in the food in parts per million.

Patients who understand the need to adhere to a gluten-free diet and the implications of not adhering to it are generally more compliant. Thus, patients need to be strongly educated that they need to adhere to a gluten-free diet and that nonadherence can cause further damage to the gut and can pose a higher risk of malignancy. Even though patients are usually concerned about the cost of gluten-free food and worry about adherence to the diet, these factors do not generally limit diet adherence.⁶⁸ All patients diagnosed with celiac disease should meet with a registered dietitian to discuss diet options based on their food preferences and to better address all their concerns.

With increasing awareness of celiac disease and with increasing numbers of patients being diagnosed with it, the food industry has recognized the need to produce gluten-free items. There are now plenty of food products available for these patients, who no longer have to forgo cakes, cookies, and other such items. Table 2 lists some common foods that patients with celiac disease can consume.

Nutritional supplements for some

If anemia is due purely to iron deficiency, it may resolve after starting a gluten-free diet, and no additional supplementation may be needed. However, if it is due to a combination of iron plus folate or vitamin B₁₂ deficiency, then folate, vitamin B₁₂, or both should be given.

In addition, if the patient is found to have a deficiency of vitamin D, then a vitamin D supplement should be given.⁶⁹ At the time of diagnosis, all patients with celiac disease should be screened for deficiencies of vitamins A, B₁₂, D, E, and K, as well as copper, zinc, folic acid, and iron.

Follow-up at 3 to 6 months

A follow-up visit should be scheduled for 3 to 6 months after the diagnosis and after that on an annual basis, and many of the abnormal laboratory tests will need to be repeated.

If intestinal or extraintestinal symptoms or nutrient deficiencies persist, then the patient’s adherence to the gluten-free diet needs to be checked. Adherence to a gluten-free diet can be assessed by checking for serologic markers of celiac disease. A decrease in baseline values can be seen within a few months of starting the diet.⁷⁰ Failure of serologic markers to decrease by the end of 1 year of a gluten-free diet usually indicates gluten contamination.⁷¹ If adherence is confirmed (ie, if baseline values fall) but symptoms persist, then further workup needs to be done to find the cause of refractory disease.

Skin lesions should also respond to a gluten-free diet

The first and foremost therapy for the skin lesions in dermatitis herpetiformis is the same as that for the intestinal manifestations in celiac disease, ie, adherence to a gluten-free diet. Soon after patients begin a gluten-free diet, the itching around the skin lesions goes away, and over time, most patients have complete resolution of the skin manifestations.

Dapsone is also frequently used to treat dermatitis herpetiformis if there is an incomplete response to a gluten-free diet or as an adjunct to diet to treat the pruritus. Patients often have a good response to dapsone.⁷²

The recommended starting dosage is 100 to 200 mg a day, and a response is usually seen within a few days. If the symptoms do not improve, the dose can be increased. Once the lesions resolve, the dose can be tapered and patients may not require any further medication. In some cases, patients may need to be chronically maintained on the lowest dose possible, due to the side effects of the drug.³

Dapsone is associated with significant adverse effects. Methemoglobinemia is the most common and is seen particularly in dosages exceeding 200 mg a day. Hemolytic anemia, another common adverse effect, is seen with dosages of more than 100 mg a day. Patients with a deficiency of glucose-6-phosphate dehydrogenase (G6PD) are at increased risk of hemolysis, and screening for G6PD deficiency is usually done before starting dapsone. Other rare adverse effects of dapsone include agranulocytosis, peripheral neuropathy, psychosis,⁷³ pancreatitis, cholestatic jaundice, bullous and exfoliative dermatitis, Stevens-Johnson syndrome, toxic epidermal necrolysis, nephrotic syndrome, and renal papillary necrosis.

Besides testing for G6PD deficiency, a complete blood cell count, a reticulocyte count, a hepatic function panel, renal function tests, and urinalysis should be done before starting dapsone therapy and repeated while on therapy. The complete blood cell count and reticulocyte count should be checked weekly for the first month, twice a month for the next 2 months, and then once every 3 months. Liver and renal function tests are to be done once every 3 months.⁷⁴

NOVEL THERAPIES BEING TESTED

Research is under way for other treatments for celiac disease besides a gluten-free diet.

Larazotide (Alba Therapeutics, Baltimore, MD) is being tested in a randomized, placebo-controlled trial. Early results indicate that it is effective in controlling both gastrointestinal and nongastrointestinal symptoms of celiac disease, but it still has to undergo phase 3 clinical trials.

Sorghum is a grain commonly used in Asia and Africa. The gluten in sorghum is different from that in wheat and is not immunogenic. In a small case series in patients with known celiac disease, sorghum did not induce diarrhea or change in levels of antibodies to tissue transglutaminase.⁷⁵

Nonimmunogenic wheat that does not contain the immunogenic gluten is being developed.

Oral enzyme supplements called glutenases are being developed. Glutenases can cleave gluten, particularly the proline and glutamine residues that make gluten resistant to degradation by gastric, pancreatic, and intestinal brush border proteases. A phase 2 trial of one of these oral enzyme supplements showed that it appeared to attenuate mucosal injury in patients with biopsy-proven celiac disease.⁷⁶

These novel therapies look promising, but for now the best treatment is lifelong adherence to the gluten-free diet.

NONRESPONSIVE AND REFRACTORY CELIAC DISEASE

Celiac disease is considered nonresponsive if its symptoms or laboratory abnormalities persist after the patient is on a gluten-free diet for 6 to 12 months. It is considered refractory if symptoms persist or recur along with villous atrophy despite adherence to the diet for more than 12 months in the absence of other causes of the symptoms. Refractory celiac disease can be further classified either as type 1 if there are typical intraepithelial lymphocytes, or as type 2 if there are atypical intraepithelial lymphocytes.

Celiac disease is nonresponsive in about 10% to 19% of cases,⁷⁶ and it is refractory in 1% to 2%.⁷⁷

Managing nonresponsive celiac disease

The first step in managing a patient with nonresponsive celiac disease is to confirm the diagnosis by reviewing the serologic tests and the biopsy samples from the time of diagnosis. If celiac disease is confirmed, then one should re-evaluate for gluten ingestion, the most common cause of nonresponsiveness.⁷⁸ If strict adherence is confirmed, then check for other causes of symptoms such as lactose or fructose intolerance. If no other cause is found, then repeat the duodenal biopsies with flow cytometry to look for CD3 and CD8 expression in T cells in the small-bowel mucosa.⁷⁹ Presence or absence of villous atrophy can point to possible other causes of malabsorption including pancreatic insufficiency, small intestinal bowel overgrowth, and microscopic colitis.

Managing refractory celiac disease

Traditionally, corticosteroids have been shown to be beneficial in alleviating symptoms in patients with refractory celiac disease but do not improve the histologic findings.⁸⁰ Because of the adverse effects associated with long-term corticosteroid use, azathioprine has been successfully used to maintain remission of the disease after induction with corticosteroids in patients with type 1 refractory celiac disease.⁸¹

Cladribine, a chemotherapeutic agent used to treat hairy cell leukemia, has shown some benefit in treating type 2 refractory celiac disease.⁸²

In type 2 refractory celiac disease, use of an immunomodulator agent carries an increased risk of transformation to lymphoma.

Because of the lack of a satisfactory response to the agents available so far to treat refractory celiac disease, more treatment options acting at the molecular level are being explored.

NONCELIAC GLUTEN SENSITIVITY DISORDER

Nonceliac gluten sensitivity disorder is an evolving concept. The clinical presentation of this disorder is similar to celiac disease in that patients may have diarrhea or other extra­intestinal symptoms when on a regular diet and have resolution of symptoms on a gluten-free diet. But unlike celiac disease, there is no serologic or histologic evidence of celiac disease even when patients are on a regular diet.

One of every 17 patients who presents with clinical features suggestive of celiac disease is found to have nonceliac gluten sensitivity disorder, not celiac disease.⁸³ In contrast to celiac disease, in which the adaptive immune system is thought to contribute to the disease process, in nonceliac gluten sensitivity disorder the innate immune system is believed to play the dominant role,⁸⁴ but the exact pathogenesis of the disease is still unclear.

The diagnosis of nonceliac gluten sensitivity disorder is one of exclusion. Celiac disease needs to be ruled out by serologic testing and by duodenal biopsy while the patient is on a regular diet, and then a trial of a gluten-free diet needs to be done to confirm resolution of symptoms before the diagnosis of nonceliac gluten sensitivity disorder can be established.

As with celiac disease, the treatment involves adhering to a gluten-free diet, but it is still not known if patients need to stay on it for the rest of their life, or if they will be able to tolerate gluten-containing products after a few years.