Features, Diagnosis, and Treatment of Nonalcoholic Fatty Liver Disease

Features, Diagnosis, and Treatment of Nonalcoholic Fatty Liver DiseaseCLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2012;10:837–858
Features, Diagnosis, and Treatment of Nonalcoholic Fatty Liver Disease
DAWN M. TORRES,* CHRISTOPHER D. WILLIAMS,‡ and STEPHEN A. HARRISON‡
*Division of Gastroenterology, Department of Medicine, Walter Reed National Military Medical Center, Bethesda, Maryland; and ‡Division of Gastroenterology, Department of Medicine, San Antonio Military Medical Center, Fort Sam Houston, Texas
As the global incidence of obesity has increased, nonalco- holic fatty liver disease (NAFLD) has become a worldwide health concern. NAFLD occurs in children and adults of all ethnicities and includes isolated fatty liver and nonalco- holic steatohepatitis (NASH). Patients with NASH are at risk for developing cirrhosis, hepatic decompensation, and hepatocellular carcinoma and have increased all-cause mor- tality. NAFLD is associated with a variety of clinical condi- tions and is an independent risk factor for hepatocellular carcinoma. The pathogenesis of NAFLD and the specific steps that lead to NASH and advanced fibrosis are not fully understood, although researchers have found that a com- bination of environmental, genetic, and metabolic factors lead to advanced disease. There have been improvements in noninvasive radiographic methods to diagnose NAFLD, es- pecially for advanced disease. However, liver biopsy is still the standard method of diagnosis for NASH. There are many challenges to treating patients with NASH, and no therapies have been approved by the U.S. Food and Drug Administration; multimodal approaches are being devel- oped and becoming the standard of care. We review patho- genesis and treatment approaches for the West’s largest liver-related public health concern.
Keywords: Diagnosis; Epidemiology; Cancer Risk; Mechanisms.
Nonalcoholic fatty liver disease (NAFLD) has become a worldwide public health concern particularly for patients
ho meet criteria for nonalcoholic steatohepatitis (NASH), ith its adherent risks for progression to cirrhosis. Evidence ow links NAFLD to cardiovascular disease (CVD), diabetes, ype 2 diabetes mellitus, obstructive sleep apnea (OSA), colonic denomas, hyperuricemia, vitamin D deficiency, hyperferritine- ia, pancreatic steatosis, hypothyroidism, and polycystic ovar-
an syndrome (PCOS).1–3
Recent data have led to an improved understanding of the natural history of NAFLD. Although NAFLD is associated with increased all-cause mortality, isolated fatty liver (IFL) behaves differently from NASH, with no increase in liver-related mor- tality and minimal risk for disease progression. NASH with fibrosis progresses at a faster rate than NASH without fibrosis. Although evidence clearly supports the development of hepa- tocellular carcinoma (HCC) in patients with NASH cirrhosis, data now suggest that HCC can also occur in NASH patients without advanced fibrosis.
The pathogenesis of NAFLD has not been completely eluci- dated, although our understanding of the complex multifac- eted processes that lead to the histopathologic features of
NASH is improving. Insulin resistance (IR) contributes, but it is
evident that environmental and genetic factors also play a role in the development of necroinflammation and subsequent fi- brosis. The dogma of sequential progression of disease from IFL to NASH to cirrhosis is being questioned. A comprehensive understanding of the steps that lead to steatohepatitis and fibrosis is a focus of active research, and this review provides for summary of current understanding of these processes.
The detection of hepatic steatosis is relatively simple, whereas the diagnosis of NASH still requires a liver biopsy. Noninvasive tests to identify patients at greatest risk for pro- gression of disease are sorely needed. However, diagnostic chal- lenges remain, including defining optimal imaging modalities and the development of a highly accurate and readily available noninvasive biomarker marker system for detecting NASH and NASH with advanced fibrosis.
Therapeutic approaches may be grouped into lifestyle mod- ification, surgical weight loss, and pharmacotherapy. Although lifestyle change remains the cornerstone of treatment, no single treatment has proved optimal, and combination therapies may be the future of NASH treatment. Future directions that focus on improving the diagnosis and treatment of this important disease entity are certain to impact global health.
Abbreviations used in this paper: ALT, alanine aminotransferase; APOC3, apolipoprotein C3; ARFI, acoustic radiation force impulse elastography; AST, aspartate aminotransferase; ATF6, activating tran- scription factor-6; AUROC, area under the receiver operating curve; BMI, body mass index; CT, computed tomography; CVD, cardiovascular disease; ELF, European Liver Fibrosis; ELISA, enzyme-linked immu- nosorbent assay; EMT, epithelial-to-mesenchymal transition; ER, endo- plasmic reticulum; FFA, free fatty acids; FXR, farnesoid X receptor; HCC, hepatocellular carcinoma; HDL, high-density lipoprotein; IFL, iso- lated fatty liver; IHTG, intrahepatic triglyceride content; IL, interleukin; IQR, interquartile range; IR, insulin resistance; IRE1, inositol requiring enzyme-1; JNK, c-Jun N-terminal kinase; LAGB, laparoscopic adjust- able gastric banding; LPS, lipopolysaccharide; LRH-1, liver receptor homolog-1; LSM, liver stiffness measurement; LXR, liver X receptor; MRI, magnetic resonance imaging; NAFLD, nonalcoholic fatty liver disease; NAS, NAFLD Activity Score; NASH, nonalcoholic steatohepa- titis; NKT, natural killer T; NPV, negative predictive value; OSA, obstruc- tive sleep apnea; PCOS, polycystic ovarian syndrome; PPAR, peroxi- some proliferator-activated receptor; PPV, positive predictive value; PTX, pentoxifylline; RYGB, Roux-en-Y gastric bypass; SFAs, saturated fatty acids; SHP, short heterodimer partner; SNPs, single nucleotide polymorphisms; SREBP, sterol regulatory element binding protein-1; TE, transient elastography; TGR5, G-protein coupled receptor; TNF, tumor necrosis factor; TZD, thiazolidinedione; UDCA, ursodeoxycholic acid; US, ultrasound.
© 2012 by the AGA Institute 1542-3565/$36.00
s s b ( r a
2 p t t
p s N h a t
l t i d
838 TORRES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 8
Definitions NAFLD is defined as macrovesicular fat accumulation
in more than 5% of hepatocytes. Fat accumulation usually begins in acinar zone 3 but may occupy the entire acinus. IFL, as opposed to NASH, is the term used when no necroinflam- matory or fibrotic changes are present on liver biopsy. NAFLD with or without NASH is indistinguishable from alcohol-in- duced liver damage on liver biopsy.
Minimal histologic criteria for the diagnosis of adult NASH include macrovesicular steatosis, hepatocyte ballooning degen- eration, and lobular inflammation.4 Lobular inflammation is typically mild, with a mixed inflammatory cell infiltrate. Mal- lory–Denk bodies, vacuolated nuclei in periportal hepatocytes, iron deposition, ductular reaction, megamitochondria, lobular lipogranulomas, periodic acid-Schiff– diastase–resistant Kupffer cells, and acinar zone 3 perisinusoidal/pericellular fibrosis may also be seen.5 Portal and periportal fibrosis may appear with
ASH progression, followed by bridging fibrosis and cirrhosis. teatosis may be absent in cases of bridging fibrosis or cirrhosis. istinctive histologic features of pediatric NAFLD include por-
al-based (acinar zone 1) chronic inflammation and fibrosis, ore severe steatosis, and decreased incidence of ballooning
epatocyte degeneration and Mallory–Denk bodies.6
NAFLD is considered by many to be the hepatic manifesta- tion of metabolic syndrome, defined as the presence of 3 or more of the following: (1) abdominal obesity (waist circumfer- ence 102 cm in men, 88 cm in women), (2) hypertriglycer- idemia (150 mg/dL), (3) low high-density lipoprotein (HDL) levels (40 mg/dL in men, 50 mg/dL in women), (4) hyper- tension (130/80 mm/Hg), and (5) high fasting glucose levels (110 mg/dL).7 Other conditions that result in hepatic steato- is include total parenteral nutrition, rapid weight loss, acute tarvation, abdominal surgery (extensive small bowel resection, iliopancreatic diversion, jejunoileal bypass), drugs or toxins amiodarone, tamoxifen, glucocorticoids, estrogen, antiretrovi- al agents, tetracycline), abetalipoproteinemia, lipodystrophy, nd Wilson’s disease.
Epidemiology Obesity and diabetes are global health concerns, with
recent estimates suggesting an overall obesity prevalence of 33.8% and diabetes prevalence among middle-aged adults of 10.6% in the United States.8 Similarly, the prevalence estimates for NAFLD are rising, now estimated to be from 2.8%– 46%, depending on the study population and diagnostic method used.9,10 Autopsy studies have placed the prevalence of NASH at
.7% among lean patients to 18.5% among markedly obese atients.11 The prevalence of NAFLD and NASH in obese pa- ients undergoing bariatric surgery has been estimated to be up o 91% and 37%, respectively.12 A recently published prospective
cohort study that used ultrasound and liver biopsy determined the prevalence of NAFLD in asymptomatic middle-aged pa- tients in the United States to be 46% and the prevalence of NASH to be 12.2%.10
An ethnic variation in the distribution of NAFLD has also been suggested, with Hispanics having the highest prevalence (45%–58%), followed by whites (33%– 44%) and blacks (24%– 35%)10,13 (Figure 1). Reasons for this ethnic variation have not
een completely defined, although both lifestyle and genetic
redisposition are important. A recent study by the NASH t
Clinical Research Network found that Hispanics with NASH tended to be younger, were less active, and ingested more carbohydrates than whites.14 IR was not a risk factor for NASH
mong Hispanics. When controlling for obesity, more recent ata suggest that there is no significant increase in hepatic teatosis or the severity of NASH and the rates of adipose tissue nd hepatic IR are similar when comparing Hispanics with hites. The one caveat is that diabetic Hispanics had more brosis.15 NAFLD is now reaching epidemic proportions mong populations that have been considered at lower risk for his disease in the past. Chinese and Japanese population-based ohort studies report a prevalence of NAFLD of 14%–15%.16,17
Several studies have demonstrated that NAFLD and NASH are significantly more common in men.10,13,18 Women have also been shown to develop a later, presumed postmenopausal, peak of NAFLD incidence compared with men.19 This suggests that sex steroid hormone metabolism may play a role in NAFLD pathogenesis. There may be differences in lifestyle as well, because data suggest that men drink more nondiet soda than women.10 The prevalence of NAFLD and NASH in the diabetic
opulation deserves special mention. Studies to date have de- cribed a 60%–76% prevalence of NAFLD and 22% prevalence of ASH in diabetics.10,20 Patients with NAFLD and diabetes have igher mortality and a higher prevalence of CVD than nondi- betic NAFLD patients, and diabetes is an independent predic- or of having advanced fibrosis.21,22
NAFLD has also become common in the pediatric popula- tion because of a 30% prevalence of elevated body mass index (BMI) (overweight or obese) in the United States.23 In the absence of large population-based studies, an autopsy study of 742 children and adolescents 2–19 years old showed an overall NAFLD prevalence of 13%.24 A recently published retrospective ongitudinal hospital-based cohort study demonstrated that he presence of NAFLD in children is associated with a signif- cantly shorter long-term survival and a 13.8-fold higher risk of ying or requiring liver transplantation when compared with
Figure 1. Prevalence of NAFLD as assessed by 2 large prospective studies. The San Antonio study used US as a screening modality for NAFLD in a tertiary care referral center. All patients found to have NAFLD on imaging were offered a liver biopsy, thus allowing for the calculation of NASH prevalence. The Dallas Heart Study performed magnetic resonance spectroscopy to assess for NAFLD in a large urban population.
he age- and gender-matched general population.25
h n
t g
Clinical Presentation and Natural History Most patients with NAFLD, including NASH, are
asymptomatic, although fatigue, malaise, and vague right upper quadrant pain have been reported. The most common presen- tation is either an elevation in the serum aminotransaminases, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), detected through routine testing or incidental evidence of hepatic steatosis on imaging. Serum ALT and AST levels are generally only mildly elevated. A large retrospective study of NAFLD patients has shown the mean ALT and AST levels to be 83 and 62 IU/mL, respectively.26 Patients with NASH tend to
ave higher ALT and AST levels than patients with fatty liver ot meeting histopathologic criteria for NASH.10 However, se-
rum aminotransferases are normal in more than half of NAFLD patients, despite having histopathologic evidence of NASH on biopsy, and thus cannot be used alone to distinguish IFL disease from NASH.27
Gamma-glutamyltransferase level may be elevated in pa- tients with NAFLD and has been associated with advanced fibrosis28 and all-cause mortality,29 possibly more so in men.30 A ecent prospective Belgian study in 230 NAFLD patients found hat a gamma-glutamyltransferase level 5 times the upper imit of normal was associated with older age, diabetes, and
ore advanced fibrosis.31
The natural history of NAFLD is dependent on the histopa- thology at baseline (Figure 2). Numerous longitudinal studies, albeit with small overall numbers of patients, now confirm that the vast majority of patients identified as having IFL on initial
Figure 2. Natural history of both phenotypes of NAFLD including IFL, which shows minimal risk for progression to cirrhosis or in- creased mortality, and NASH, which shows increased overall mortality as well as increased risk for cirrhosis and HCC. DM, dia-
betes mellitus.
biopsy are at extremely low risk for the development of ad- vanced disease.32–36 Collectively, data suggest that there does
ot appear to be an increased risk of death among patients with FL compared with age- and gender-matched controls. An ear- ier study by Adams et al37 showed a 1.34 hazard ratio for
ortality in NAFLD patients, but this did not distinguish etween the various subtypes of NAFLD.
Once NASH has been diagnosed on liver biopsy, data suggest hat there is indeed an increased risk of death compared with the eneral population.31,33 Causes of death in order of decreasing
frequency include cardiovascular, malignancy, and liver-re- lated.31,33,34 NASH with fibrosis has been shown to carry a worse prognosis than NASH without evidence of fibrosis. Fibrosis pro- gression is associated with diabetes mellitus, severe IR, BMI, weight gain 5 kg, cigarette smoking, and rising ALT and AST levels.38,39
Progression to NASH cirrhosis has been more difficult to predict. Although the current literature lacks adequately pow- ered studies with repeat histopathology obtained over many years to several decades, data suggest that progression to cir- rhosis is a long-term process. A review of the published longi- tudinal trials on progression of disease shows an average pro- gression rate of 11% during a 15-year period.40 However, it is lear that the fibrosis progression rate is highly variable and ontingent on multiple factors, with the two most important eing obesity and type 2 diabetes mellitus.
The progression to end-stage liver disease or HCC once dvanced fibrosis or cirrhosis is present remains somewhat ague. Fortunately, a few recent studies have improved our nderstanding of the progression of disease. In a recent inter-
p H t d m
i o w fi
a c s e
m w s m o o t
1 e i o k N N g a p p
840 TORRES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 8
national, multicenter prospective observational study, Bhala et al41 evaluated 247 biopsy-confirmed NAFLD patients with well- compensated advanced fibrosis or cirrhosis. They followed these patients for a mean of 7.1 years and compared them with a group of 264 chronic hepatitis C patients with Child–Pugh class A advanced fibrosis or cirrhosis who were followed for a mean of 6.2 years. The NAFLD group developed liver-related complications in 19.4% and HCC in 2.4% vs 17.8% liver-related complications and 6.8% HCC in patients with hepatitis C virus. When adjusting for confounders such as age, sex, BMI, and diabetes, the hepatitis C virus cohort had significantly more liver-related complications. There was no difference in vascular events between groups, and overall mortality was similar be- tween groups. The 10-year survival of the NAFLD cohort was 81.5%, similar to the study by Sanyal et al,42 who found a 10-year survival of 80.9% in a smaller cohort of NAFLD Child– Pugh class A patients. When the studies assessing progression of NASH cirrhosis are averaged together, it is estimated that approximately 31% of patients with NASH cirrhosis will prog- ress to decompensation during an 8-year period.38,39,43
The development of HCC is one of increasing importance as the prevalence of NASH continues to increase. There is clearly an increased risk of developing HCC among patients with NASH cirrhosis. A review of several recent studies assessing the outcome of patients with NASH cirrhosis demonstrates that 7% will progress to HCC during 6.5 years of follow-up.41– 44
The data range from 2.4%–12.8% during 3.2–10 years.45,46 Com- pared with HCV cirrhosis, the incidence and prevalence of HCC are less. It is estimated that the yearly incidence of HCC among NASH cirrhotic patients is 2.6%–2.7%, compared with 4%– 4.7% for HCV cirrhotic patients.43,47 Overall survival after liver trans-
lantation is the same.43 Risk factors for the development of CC include age, obesity, diabetes, and iron deposition, al-
hough there is discordance in the literature in regard to gender ifference.48 A large, cross-sectional Japanese study showed a ale predominance,49 but this was not supported by a recent
large, cross-sectional study from the United States.43
A major concern is the potential for HCC to arise in non- cirrhotic NASH. Data are beginning to accumulate to suggest that this is a real and growing problem. Most of the data are in the form of case reports or small case series.42,43 When combin- ng 3 of the largest case series with a recent cross-sectional study f 87 histopathologically proven NASH cases, 36 of 138 (26%) ere found to have HCC in the setting of mild-to-moderate brosis (stage 0 –2).42,50,51 A French study also identified pa-
tients with little to no fibrosis with HCC, which appeared to occur in association with metabolic syndrome.52
Transplant Considerations Clinicians are faced with 2 separate issues related to
NAFLD and liver transplantation. First, hepatologists and transplant surgeons are faced with the frequent consideration of using harvested livers with varying degrees of steatosis for liver transplant. Transplanted steatotic livers are associated with primary graft nonfunction, poor initial function, and poorer outcomes in general.53 Steatosis is currently detected in
pproximately one-third to one-half of potential living and adaveric liver transplant donors.54 Livers with less than 30% teatosis are considered optimal, whereas severely steatotic liv-
rs with 60% fat are not typically used. Livers with interme-
diate quantities of steatosis from 30%– 60% may be assessed for use on a case-by-case basis.55
NASH is now the third most common indication for liver transplantation (behind hepatitis C and alcoholic liver disease) in the United States and is expected to become the number one indication within the next 8 –15 years.56 Unfortunately, a sig-
ificant number of NASH cirrhotic patients will never undergo iver transplantation because of comorbid conditions, particu- arly obesity.47 Transplants performed on known NASH cir-
rhotic or obese patients with cryptogenic cirrhosis have both short- and long-term considerations. NASH cirrhosis has been linked to increased 30-day transplant mortality,57 but generally survival at 1 and 3 years after transplant is similar to other indications for transplant. Malik et al58 identified a subgroup of NASH cirrhotic patients (those with age 60 years, BMI 30 kg/m2, diabetes, and hypertension) who had an increased 1-year
ortality of 50%. Unfortunately, recurrent NAFLD in patients ho undergo transplantation for NASH or cryptogenic cirrho-
is is high, occurring in 30% of patients by 1 year and the ajority of patients within 5 years. Five percent of patients go
n to develop cirrhosis within 5 years after transplant, although verall mortality is similar to patients who undergo transplan- ation for other etiologies of cirrhosis.59,60
Clinical Associations Clinically associated comorbidities of NAFLD and
NASH are numerous and include CVD, diabetes, OSA, colonic adenomas, hyperuricemia, vitamin D deficiency, hyperferritine- mia, pancreatic steatosis, hypothyroidism, and PCOS (Figure 3).
CVD is the number one cause of mortality in NAFLD. NAFLD patients share many known cardiovascular risk factors including obesity, IR, type 2 diabetes, hypertension, dyslipide- mia, and metabolic syndrome. NAFLD has been shown to be an independent risk factor for CVD, owing in part to an associa- tion with greater carotid artery intima-media thickness.61,62 A meta-analysis of 8 studies evaluating NAFLD (defined by ultra- sound or histology) as a risk factor for CVD found an overall odds ratio of 2.05 (95% confidence interval, 1.81–2.31).63
NAFLD was also found to increase the risk of developing diabetes mellitus 2- to 3-fold,61,64 although in a single-centered,
1-year follow-up study of patients with NAFLD defined by levated ALT and exclusion…