In the June issue, we covered Part 1 of Nonalcoholic fatty liver disease (NAFLD). The article went over definition, epidemiology, pathogenesis, and clinical manifestations of Nonalcoholic fatty liver disease (NAFLD). In part 2 of the article, we will continue to cover Nonalcoholic fatty liver disease (NAFLD), and we will go over diagnosis, which patients to biopsy, differential diagnosis, and screening.
DIAGNOSIS — The diagnosis of nonalcoholic fatty liver disease (NAFLD) requires all of the following :
- Demonstration of hepatic steatosis by imaging or biopsy
- Exclusion of significant alcohol consumption
- Exclusion of other causes of hepatic steatosis
In those undergoing a radiologic evaluation, radiologic findings are often sufficient to make the diagnosis if other causes of hepatic steatosis have been excluded. While not indicated for the majority of patients, a liver biopsy may be indicated if the diagnosis is not clear or to assess the degree of hepatic injury. In addition, liver biopsy is the only method currently available to differentiate nonalcoholic fatty liver (NAFL) from nonalcoholic steatohepatitis (NASH).
Laboratory tests — Laboratory tests, such as the serum aminotransferase and ferritin levels, are often abnormal in NAFLD. However, these abnormalities are neither required nor sufficient for making the diagnosis, as laboratory tests may be normal in patients with NAFLD and may be abnormal in patients with numerous other conditions.
However, laboratory testing is required to evaluate for other conditions in the differential diagnosis of hepatic steatosis.
Rule out other disorders — Differentiating NAFLD from the other items in the differential diagnosis begins with a thorough history to identify potential causes such as significant alcohol use, starvation, medication use, and pregnancy-related hepatic steatosis.
We test all patients with hepatic steatosis for hepatitis C virus infection. We also test for hepatitis A and B. We do this to both to rule out these infections in patients with elevated aminotransferases and to determine immunity to guide future immunizations. We also rule out other chronic liver diseases such as autoimmune hepatitis and hemochromatosis.
We obtain the following tests in all patients:
- Anti-hepatitis C virus antibody
- Hepatitis A IgG
- Hepatitis B surface antigen, surface antibody, and core antibody
- Plasma iron, ferritin, and total iron binding capacity
- Serum gammaglobulin level, antinuclear antibody, antismooth muscle antibody, and anti-liver/kidney microsomal antibody-1
Other disorders that should be considered based upon the patient’s history, associated symptoms, and family history include Wilson disease, thyroid disorders, celiac disease, alpha-1 antitrypsin deficiency, HELLP, and Budd-Chiari syndrome.
Radiographic examinations — Various radiologic methods can detect NAFLD, but no imaging modality is able to differentiate between the histologic subtypes of nonalcoholic fatty liver (NAFL) and nonalcoholic steatohepatitis (NASH) . Our approach in patients who have not already undergone imaging is to obtain an ultrasound. However, computed tomography (CT) and magnetic resonance imaging (MRI) can also detect hepatic steatosis.
We consider a radiographic diagnosis to be sufficient for diagnosing NAFLD if all of the following conditions are met:
- Radiographic imaging is consistent with fatty infiltration
- Other causes for the patient’s liver disease have been excluded
- The patient does not have signs or symptoms cirrhosis
- The patient is not at high risk for advanced fibrosis or cirrhosis (eg, a younger patient who does not have diabetes and has a normal serum ferritin is at lower risk for having fibrosis or cirrhosis)
If these criteria are not met, patients will typically require a liver biopsy to make the diagnosis or to assess the degree of liver injury.
Ultrasound — Ultrasonography often reveals a hyperechoic texture or a bright liver because of diffuse fatty infiltration . A meta-analysis of 49 studies with 4720 patients found that the sensitivity and specificity for ultrasound were 85 and 94 percent, respectively, when using liver biopsy as the gold standard [. However, the sensitivity appears to be decreased in patients who are morbidly obese . In a study of 187 morbidly obese patients undergoing bariatric surgery, hepatic steatosis was present histologically in 95 percent but was only detected by ultrasound in 49 percent .
CT, MRI, and magnetic resonance spectroscopy — Both CT and MRI can identify steatosis but are not sufficiently sensitive to detect inflammation or fibrosis . Magnetic resonance spectroscopy (MRS) has the advantage of being quantitative rather than qualitative or semiquantitative, but it is not widely available .
One of the difficulties in determining the sensitivity and specificity of CT and MRI for diagnosis of hepatic steatosis is that not all patients undergo confirmation by liver biopsy. In a study that did use histology as the gold standard, the sensitivity of CT scan for detecting hepatic steatosis was poor, whereas MRI had low specificity ]. It included a total of 131 patients who had a radiologic evaluation with noncontrast CT, contrast-enhanced CT, or MRI before undergoing a partial hepatectomy, usually for malignancy. The sensitivities of noncontrast CT, contrast-enhanced CT, and MRI for detecting hepatic steatosis were 33, 50, and 88 percent, respectively. The specificities were 100, 83, and 63 percent, respectively. In addition, the accuracy of noncontrast CT fell with increasing body mass index.
Unlike CT and MRI, MRS allows for quantification of hepatic fat, and may be particularly helpful in patients with small amounts of hepatic steatosis 2]. A study that compared MRS with liver biopsy in 12 patients found a close correlation between the measurement of intrahepatocellular lipid by MRS and the histologic assessment of cirrhosis (r = 0.94) . However, not all scanners have the capability of obtaining spectroscopic sequences, and it is not routinely used.
Role of liver biopsy — While liver biopsy is the gold standard for diagnosing NAFLD, in many cases a presumptive diagnosis can be made based upon the patient’s history, laboratory tests, and imaging findings, provided other disorders have been excluded. However, some patients will continue to have an unclear diagnosis following a noninvasive evaluation. In such cases, a liver biopsy is indicated.
In addition, imaging studies and laboratory tests do not reliably differentiate patients with NAFL from those with NASH, or predict the severity of liver disease ]. The only way to definitively confirm or exclude the diagnosis of NASH and to determine disease severity is with a liver biopsy ]. This information can be used to guide patient care and may motivate patients to enact lifestyle modifications. As examples, patients found to have cirrhosis will require screening for esophageal varices and hepatocellular carcinoma, whereas patients with early fibrosis may be motivated to lose weight to decrease the risk of progressing to cirrhosis.
A potentially useful non-invasive method for excluding advanced fibrosis is measurement of liver stiffness with transient elastography. However, the approach is not widely available and has not been extensively studied in NASH. Other indirect markers of cirrhosis such as the aspartate aminotransferase to platelet ratio index are also being studied to identify patients with fibrosis.
Which patients to biopsy — There is no clear consensus about which patients require a liver biopsy . We obtain a liver biopsy in patients with suspected NAFLD if the diagnosis is unclear after obtaining standard laboratory tests and hepatic imaging, if there is evidence of cirrhosis, if the patient wants to know if inflammation or fibrosis is present, or if the patient is at increased risk for advanced fibrosis or cirrhosis.
Specifically, we obtain a biopsy if the patient:
- Has peripheral stigmata of chronic liver disease (suggestive of cirrhosis)
- Has splenomegaly (suggestive of cirrhosis)
- Has cytopenias (suggestive of cirrhosis)
- Has a serum ferritin >1.5 times the upper limit of normal (suggestive of NASH and advanced fibrosis)
- Is >45 years of age with associated obesity or diabetes (increased risk of advanced fibrosis)
Histologic findings — Histologic findings in NAFLD include steatosis, inflammation, cell injury, and fibrosis. The minimum criterion for a histologic diagnosis of NAFLD is >5 percent steatotic hepatocytes in a liver tissue section . The extent of steatosis can be described as mild (5 to 33 percent of hepatocytes are steatotic), moderate (34 to 66 percent of hepatocytes), or severe (>66 percent of hepatocytes) .
Patients with NAFLD typically have macrovesicular steatosis, though mixed steatosis may also be seen . Pure microvesicular steatosis is uncommon. In adults, steatosis is typically first seen in acinar zone 3, though when severe it may occupy the entire acinus.
Patients with NAFL may have foci of lobular inflammation, mild portal inflammation, and lipogranulomas, but features of steatohepatitis (ie, hepatocellular injury and fibrosis) are absent by definition.
Patients with NASH have liver biopsy findings that may be indistinguishable from those of alcoholic steatohepatitis. A diagnosis of NASH requires the findings of steatosis, hepatocyte injury (typically ballooning degeneration), and lobular inflammation (typically in acinar zone 3). Fibrosis is not a required diagnostic feature, but may be seen.
Histologic findings of NASH include
- Hepatocyte swelling or ballooning degeneration
- Apoptotic (acidophil) bodies
- Mild lobular inflammation (acute, and less often, chronic)
- Mild chronic portal inflammation (inflammation that is severe or is disproportionate to the acinar lesions is suggestive of concurrent hepatitis C)
- Perisinusoidal collagen deposition that may result in zone 3 accentuation in a “chicken wire” pattern (related to the deposition of collagen and other extracellular matrix fibers along the sinusoids of zone 3 and around hepatocytes)
- Portal fibrosis without perisinusoidal or pericellular fibrosis
- Cirrhosis, which is typically macronodular or mixed
- Mallory-Denk bodies (previously called Mallory bodies or Mallory’s hyaline)
- Glycogenated (vacuolated) nuclei in periportal hepatocytes (rarely seen in alcoholic steatohepatitis)
- Lobular lipogranulomas
- PAS-diastase-resistant Kupffer cells
- Hepatic siderosis (typically mild) involving periportal hepatocytes or panacinar reticuloendothelial cells
As fibrosis progresses to cirrhosis, steatosis and inflammation may not be reliably identified, resulting in a diagnosis of “cryptogenic” cirrhosis . It is possible that portal fibrosis alone may represent a variant of NASH . In biopsy specimens from children, portal inflammation may be more prominent than in adults.
NASH may exist concurrently with other liver diseases, though diagnosing NASH in that setting can be difficult. As an example, patients with NASH may also have alcoholic liver disease, but there is no way to differentiate the relative contributions of the two processes from a liver biopsy]. In a series of 3581 liver biopsies from patients with various chronic liver diseases, concurrent steatohepatitis was found in 5.5 percent of patients with hepatitis C (some with significant alcohol use) . Among patients with other chronic liver diseases of nonalcoholic etiology, the prevalence ranged from 1.6 percent (autoimmune hepatitis) to 7.9 percent (alpha-1 antitrypsin deficiency). None of the patients with steatohepatitis with chronic liver disease from a cause other than hepatitis C had significant alcohol consumption.
NAFLD activity score — The NAFLD activity score (NAS) is a validated score that is used to grade disease activity in patients with NAFLD]. The NAS is the sum of the biopsy’s individual scores for steatosis (0 to 3), lobular inflammation (0 to 2), hepatocellular ballooning (0 to 2), and fibrosis (0 to 4). An NAS of 1 or 2 corresponds to NAFL, 3 to 4 corresponds to borderline NASH, and a score ≥5 corresponds to NASH.
Noninvasive assessment of hepatic fibrosis — There are now several noninvasive methods to detect fibrosis in patients with liver disease. One of the scores, the NAFLD fibrosis score, is specific to NAFLD. The score takes into account the patient’s age, body mass index, hyperglycemia, aminotransferase levels, platelet count, and albumin. Studies suggest that higher NAFLD fibrosis scores may be associated with increased mortality from cardiovascular disease.
Alternative causes of hepatic steatosis — There are multiple causes of hepatic steatosis that should be considered in a patient with suspected nonalcoholic fatty liver disease (NAFLD). Causes of hepatic steatosis in addition to NAFLD include
- Alcoholic liver disease
- Hepatitis C (particularly genotype 3)
- Wilson disease
- Parenteral nutrition
- Medications (amiodarone, methotrexate, tamoxifen, glucocorticoids, valproate, anti-retroviral agents for HIV)
- Reye syndrome
- Acute fatty liver of pregnancy
- HELLP (hemolytic anemia, elevated liver enzymes, low platelet count) syndrome
- Inborn errors of metabolism (LCAT deficiency, cholesterol ester storage disease, Wolman disease)
Significant alcohol consumption — Several definitions have been proposed for what constitutes significant alcohol consumption . We define significant alcohol consumption as an average consumption of >210 grams of alcohol per week in men or >140 grams of alcohol per week in women over at least a two-year period, a definition that is consistent with a 2012 joint guideline from the American Gastroenterological Association, the American Association for the Study of Liver Diseases, and the American College of Gastroenterology.
A standard drink in the United States (12 oz [360 mL] of beer, 5 oz [150 mL] of wine, 1.5 oz [45 mL] of 80-proof spirits) contains approximately 14 grams of alcohol , so the limits above roughly translate to >15 drinks per week for men and >10 drinks per week for women.
One finding that suggests alcoholic fatty liver disease rather than NAFLD is an aspartate aminotransferase (AST) to alanine aminotransferase (ALT) ratio >2 (it is typically <1 in patients with NAFLD).
The alcoholic liver disease to NAFLD index (ANI) is a model that has been developed to predict the probability that steatohepatitis is due to alcoholic liver disease . The model is based upon aminotransferase levels, mean corpuscular volume (MCV), body mass index (BMI), and sex:
ANI = -58.5 + 0.637 (MCV) + 3.91 (AST/ALT) – 0.406 (BMI) + 6.35 for men
An ANI greater than zero favors a diagnosis of alcoholic liver disease, whereas an ANI less than zero favors a diagnosis of NAFLD. The probability of the patient having alcoholic liver disease rather than NAFLD is then calculated using the value obtained for the ANI:
Probability = eANI/(1+eANI)
The ability of the model to accurately categorize patients ranged from good to excellent in validation cohorts .
SCREENING — One issue that arises is whether to screen patients for nonalcoholic fatty liver disease if they are at increased risk because of an associated condition such as diabetes or obesity. Currently, the American Association for the Study of Liver Diseases guidelines do not recommend screening because there are uncertainties around which diagnostic test to use (since liver enzyme levels may be normal in patients with NAFLD), how to treat NAFLD if discovered, and whether screening is cost-effective.
By Harinath Sheela, MD
Harinath Sheela, MD moved to Orlando, Florida after finishing his fellowship in gastroenterology at Yale University School of Medicine, one of the finest programs in the country. During his training he spent significant amount of time in basic and clinical research and has published articles in gastroenterology literature. His interests include Inflammatory Bowel Diseases (IBD), Irritable Bowel Syndrome (IBS), Hepatitis B, Hepatitis C, Metabolic and other liver disorders. He is a member of the American Gastroenterological Association (AGA), the American Society for Gastrointestinal Endoscopy (ASGE) and the American Association for the Study of Liver Diseases (AASLD) and Crohn’s Colitis foundation (CCF). Dr. Sheela is a Clinical Assistant Professor at the University of Central Florida School of Medicine. He is also a teaching attending physician at Florida Hospital Internal Medcine Residency and Family Practice Residence (MD and DO) programs.