Nonalcoholic Fatty Liver Disease: Making the Diagnosis

Watch a video presentation of this article

Abbreviations

ALT

alanine aminotransferase

APRI

AST to Platelet Ratio Index

AST

aspartate aminotransferase

AUROC

area under the receiver operating characteristic curve

BARD, BMI

body mass index

CK‐18

cytokeratin‐18

CRN

Clinical Research Network

FIB‐4

Fibrosis‐4

GGT

Gamma‐glutamyl transferase

HS

hepatic steatosis

MRE

magnetic resonance elastography

NAFLD

nonalcoholic fatty liver disease

NAS

NAFLD activity score

NASH

nonalcoholic steatohepatitis

NFS

NAFLD fibrosis score

NICE

National Institute for Health Care Excellence

NPV

negative predictive value

TE

transient elastography

VCTE

vibration‐controlled transient elastography

Clinical Consult

A 35‐year‐old man with diabetes and a body mass index (BMI) of 40 was found to have increased echogenicity consistent with fatty liver on ultrasound performed for right upper quadrant pain. The pain is now resolved, and liver blood tests are normal. What is the differential diagnosis? What further testing, if any, is recommended?

Making the Diagnosis of NAFLD

When imaging findings consistent with hepatic steatosis (HS) are noted, the diagnosis of nonalcoholic fatty liver disease (NAFLD) should be strongly suspected in the presence of obesity and diabetes mellitus. The global prevalence rate of NAFLD is estimated to be approximately 25%, with a recent systematic review finding NAFLD to be prevalent in 58% of individuals with diabetes. Although the prevalence is high, secondary causes of hepatic fat accumulation must be ruled out to make the diagnosis of NAFLD.1, 2

Figure 1 reviews the diagnostic work‐up recommended to investigate causes of HS seen on imaging.2, 3, 4, 5

Fig 1.

Evaluation of NAFLD in an asymptomatic patient with normal liver function tests. Suggested diagnostic algorithm to investigate causes of HS seen on imaging. In those individuals with risk factors for fatty liver disease, a stepwise approach using noninvasive testing to screen for the presence of NASH and liver fibrosis can identify patients at highest risk for complications of HS seen on imaging.

Secondary Causes of Hepatic Steatosis

HS characterized by excessive fat deposition in the liver is the initial response to excessive drinking in >90% of cases. It is the major secondary cause of fatty liver and occurs due to accelerated hepatic lipogenesis, increased fatty acid transport into the liver from the plasma, defective secretion of lipoproteins (e.g., very low‐density lipoproteins) from the liver into the plasma, and reduced mitochondrial fatty acid oxidation.6 Other competing causes of HS, including hepatitis C, Wilson's disease, medications (such as steroids, antiretroviral therapy, amiodarone, and tamoxifen) and lipid disorder, should be excluded before making a diagnosis of NAFLD.2

NAFLD is also associated with increased serum ferritin levels, which is indicative of underlying inflammation and insulin resistance. In these cases, a transferrin saturation of <45% excludes hemochromatosis.4 In addition, significant titers of autoantibodies are frequently encountered in patients with NAFLD. A study of well‐characterized individuals with NAFLD found that nearly 20% had antinuclear antibody >1:160 and/or anti–smooth muscle antibody >1:40 in the absence of autoimmune hepatitis. Although a liver biopsy may be needed to definitively rule out coexisting autoimmune hepatitis in the right clinical scenario, these autoantibodies were associated with more advanced histological features of NASH in patients with NAFLD alone.7

Beyond Simple Steatosis

Once the diagnosis of NAFLD is made, it is important to differentiate between simple steatosis and nonalcoholic steatohepatitis (NASH) for prognosis. Although invasive, liver biopsy remains the gold standard for the diagnosis of NAFLD and provides information about HS, hepatocellular inflammation, and fibrosis. NAFLD activity score (NAS) is a well‐established scoring system for the histological assessment of NAFLD. The NAS provides a composite score based on the degree of steatosis, lobular inflammation, hepatocyte ballooning, and fibrosis. A score >5 suggests probable or definite NASH, and <3 indicates that NASH is unlikely.4 Risks of liver biopsy include risk for death in 1 of 10,000, which limits the routine use of liver biopsy for diagnosis of NASH. Other major complications are bleeding requiring transfusion, pneumothorax, hemothorax, and perforation of another organ.8

Noninvasive Diagnosis of NASH

In those with HS and risk factors for NASH and/or liver fibrosis, well‐established noninvasive tests can further risk‐stratify patients with the goal of targeting those who would benefit from liver biopsy, as well as aggressive management of their metabolic syndrome. Currently, the majority of patients with NAFLD can be diagnosed and staged using noninvasive strategies.4, 5, 9

Elevated Alanine Aminotransferase for Diagnosis of NASH

Often, a clinician’s main indicator to order further diagnostic work‐up for liver injury is abnormal liver transaminases, such as an elevated alanine aminotransferase (ALT) level. Unfortunately, population studies demonstrate that current normal ranges of liver enzymes are too high, which might miss a significant number of individuals with underlying liver disease, leading to missed opportunities for intervention.10 In NAFLD, the diagnostic utility of ALT activity remains insufficient with a sensitivity of 45%.3 Transaminase levels are within normal limits in approximately 25% to 50% of patients with NAFLD and only mildly elevated in the remaining 50% to 75% of patients. In fact, the entire histological spectrum of NAFLD from macrovesicular steatosis to cirrhosis may exist without elevation of ALT levels.

Biomarkers for Diagnosis of NASH

Potential biomarkers for the diagnosis of NASH have been identified. They typically represent the key mechanisms believed to be involved in NASH pathogenesis, such as inflammation, oxidative stress, apoptosis, and insulin resistance.5 One such marker is cleaved cytokeratin‐18 (CK‐18) fragment, which is released during apoptosis and is significantly elevated in patients with biopsy‐proven NASH. A study conducted on a morbidly obese population demonstrated that the CK‐18 levels decreased significantly after bariatric surgery in patients with NASH.5 Although biomarkers such as CK‐18 may aid in the diagnosis of NASH versus simple steatosis, their performance alone has not been validated in prospective studies, and they are not available commercially.9

Predictive Models for the Diagnosis of NASH

Instead of biomarkers alone, combining clinical features with serum markers may lead to a more accurate prediction of NASH. It is well established that the risk for NAFLD and NASH increases in the presence of metabolic syndrome. In addition, with increasing the number of metabolic risk factors, there is a higher risk for progressive liver disease.2 Hence metabolic risk factor profiling can be used to identify patients with NASH until definitive diagnostic tests are available.4 Capitalizing on the diagnostic accuracy of both biomarkers and metabolic profiling has led to the development of several predictive models to determine the presence NASH. Examples include the NASHTest, NASH Clinical Research Network (CRN) model, National Institute for Health Care Excellence (NICE) model, NAFLD diagnostic panel, and oxNASH risk score. As with biomarkers, these models have demonstrated accuracy in predicting the presence of NASH but have not been externally validated in prospective cohort studies (Table 1).9, 11

Table 1.

Noninvasive Tests to Determine the Presence of NASH

Test	Diagnostic Ability (AUROC)	Comments
CK‐18 biomarker	0.82	Breakdown product during apoptosis of hepatocytes. Modest sensitivity/specificity in multiethnic cohort (58%‐68%). Not commercially available alone
NASHTest	0.78	Components include age, sex, height, weight, cholesterol, triglycerides, AST, ALT, bilirubin, haptoglobin α2‐macroglobulin, apolipoprotein A1
NASH CRN model	0.79	Components include AST level, ALT level, AST/ALT ratio, demographics (age, race, gender, and ethnicity), comorbidities (hypertension, type 2 diabetes, BMI, waist circumference, waist/hip ratio, and acanthosis nigricans), and other laboratory tests
NICE model	0.83	Components include metabolic syndrome, ALT, and CK‐18
NAFLD diagnostic panel	0.81	Based on diabetes, gender, BMI, triglycerides, M30 (CK‐18 fragments as a marker of apoptosis), and M65 plus M30 (total CK‐18 and CK‐18 fragments as a marker of necrosis)
OxNASH risk score	0.79	Score is calculated from age, BMI, AST level, and the ratio of 13‐hydroxy octadecadienoic acid to linoleic acid

Identifying Liver Fibrosis in NAFLD

Identifying those with fibrosis and staging the level of fibrosis is key in determining prognosis and guiding management.12 This is especially true for those individuals with advanced fibrosis who are at increased risk for cirrhosis‐related complications who need early, aggressive management.

Similar to predicting the presence of NASH, generic and disease‐specific models have been developed to predict liver fibrosis (Table 2).2, 11, 13 An example of a disease‐specific model is the NAFLD fibrosis score (NFS). NFS is a validated scoring system for identifying advanced fibrosis in patients with NAFLD. An example of a generic model is the Fibrosis‐4 (FIB‐4) index, which was originally developed for staging liver fibrosis in hepatitis C virus infection but also has an area under the receiver operating characteristic curve (AUROC) of 0.84 for diagnosing advanced fibrosis.11 Table 2 summarizes these models.

Table 2.

Noninvasive Tests for Detection of Liver Fibrosis

Test	Diagnostic Ability (AUROC)	Comments
Enhanced liver fibrosis panel	0.87	Detects markers of matrix turnover, which includes tissue inhibitor metalloproteinase 1, N‐terminal propeptide of type III procollagen, and hyaluronic acid
Fibrometer	0.82	Includes ALT, AST, GGT, platelets, prothrombin time index, α2‐macroglobulin, hyaluronic acid, ferritin, glucose, and urea
FibroTest	0.81	Components include age, sex, bilirubin, GGT, haptoglobin, α2‐macroglobulin, and apolipoprotein A1
BARD score	0.81	Components include BMI, AST/ALT ratio, and diabetes
NFS	0.84	Validated scoring system; components include age, diabetes, BMI, AST, ALT, platelets, and albumin
FIB‐4 index	0.84	Reliably excludes advanced fibrosis because of high NPV. Components include age, AST, and platelets
APRI	0.67	Initially developed for use in hepatitis C virus. Not specific for NAFLD
VCTE	0.83‐0.95	Results may be invalid in obese patients (BMI >35 kg/m2); hence a FibroScan XL probe is developed to overcome this problem
MRE	0.92	Widespread clinical adoption is limited because of its high cost and low availability

It is important to note that clinical history alone can help identify high‐risk patients. For example, presence of obesity, type 2 diabetes, age older than 45 years, an elevated aspartate aminotransferase (AST)/ALT ratio, hypertension, and hyperlipidemia increase the risk for progressive fibrosis and cirrhosis. Furthermore, the severity of NAFLD correlates directly with the severity of the metabolic syndrome, and the presence of more metabolic risk factors is associated with increased risk for NASH and advanced fibrosis. Therefore, some would advocate using these clinical features alone to triage patients who need further screening for NASH and NASH‐related fibrosis.3

Radiological Assessment of Liver Fibrosis

Vibration‐controlled transient elastography (VCTE) or FibroScan is an ultrasound‐based, noninvasive method of measurement of liver stiffness due to deposition of fibrous tissue in hepatic parenchyma. It has been shown to perform better than a number of noninvasive scoring systems in the staging of fibrosis with a high negative predictive value (NPV) for greater than stage 3 fibrosis according to studies. Hence a low liver stiffness measure can reliably exclude advanced fibrosis.9

A magnetic resonance equivalent of transient elastography (TE) has recently demonstrated excellent diagnostic accuracy with sensitivity and specificity of 98% and 99%, respectively, for detecting all grades of fibrosis. Magnetic resonance elastography (MRE)‐measured hepatic stiffness has the potential to identify NASH before fibrosis onset and has better diagnostic accuracy than VCTE in both obese and nonobese patients.13

A recent study that compared various risk scores and elastography (MRE and TE) against liver histology showed that NFS and FIB‐4 were better than other indices, such as BARD score, AST to Platelet Ratio Index (APRI), and AST/ALT ratio, and as good as MRE for predicting advanced fibrosis in patients with biopsy‐proven NAFLD.2

Conclusion

The finding of HS in individuals with the clinical features of metabolic syndrome warrants further evaluation. Physicians should be vigilant about the possibility of NASH, as well as advanced stages of NAFLD, despite normal ALT levels, particularly in those with features of metabolic syndrome. We suggest that noninvasive tests should be undertaken initially to rule out the presence of NASH and advanced fibrosis, and in case of indeterminate results, a liver biopsy should be performed.