Fatty Liver Disease -- A Major Cause of Obesity-Related Morbidity and Mortality

Introduction
A number of presentations at Digestive Disease Week (DDW) 2005 addressed important advances in our understanding of the hepatic consequences of obesity -- nonalcoholic fatty liver disease (NAFLD). NAFLD is the term used to describe a spectrum of disorders characterized by macrovesicular steatosis that occur in the absence of consumption of alcohol in amounts considered to be harmful to the liver. Because the likelihood of having NAFLD is directly proportional to body weight, given the increasing prevalence of obesity, NAFLD is an important public health problem. Sustained liver injury will lead to progressive fibrosis and cirrhosis in 10% to 25% of affected individuals.[1]

Natural History
There are 2 recognized histologic patterns of NAFLD: fatty liver alone and nonalcoholic steatohepatitis (NASH). The latter represents a shift from simple steatosis to an inflammatory component. NASH is described by a grading and staging system -- the histologic grade indicates the activity of the inflammatory lesion, whereas the stage reflects the progressive degree of fibrosis. In 879 patients with NASH studied by Shoji and colleagues,[2] progression was most rapid in older, female subjects. The steatosis grade was not associated with histologic activity and staging. To further assess the natural history of progression and pace of histologic change, Loomba and colleagues[3] queried the National Institutes of Health pathology database for patients with NASH with more than 2 liver biopsies, taken 3 years apart, with no intervening treatment. They documented that histologic features of disease activity in NASH (steatosis, ballooning, inflammation, and cell injury) can improve without specific therapy. However, fibrosis rarely improved on its own. In a prospective cohort study, Yatsuji and colleagues[4] described the outcome of 205 Japanese patients with NASH; 64 patients had advanced fibrosis. During a mean follow-up of 33 months, 10 patients developed liver-related morbidity, including hepatocellular carcinoma (HCC). Because HCC is the most common cause of death in patients with NASH with advanced fibrosis, surveillance for HCC must be considered.

Clinical Features
NAFLD is the hepatic component of the metabolic syndrome -- obesity, type 2 diabetes mellitus, insulin resistance, dyslipidemia, and hypertension.

In a cross-sectional, prospective study of 4,401 Japanese subjects, Hamaguchi and colleagues[5] noted that the metabolic syndrome was a strong predictor of new-onset NAFLD. The metabolic syndrome also seems to have an impact on the clinical course. Coexistence of the metabolic syndrome negatively affects overall quality of life; NAFLD patients with this syndrome had significantly lower cumulative physical and mental health scores.[6] Similarly, because depression and anxiety have been associated with insulin resistance and inflammatory states, Elwing and colleagues[7] hypothesized that these psychiatric disorders would be more prevalent in patients with NASH and would predict more severe histologic findings. Lifetime rates of these disorders -- 56% and 50%, respectively -- were significantly higher in NASH subjects than in controls; their onset preceded the diagnosis of NASH by about 20 years. Each psychiatric disorder was associated with greater age-adjusted histologic severity; these may serve as modifiable risk factors.

Screening and Diagnosis
NAFLD may not be easily or consistently recognized and there are scant data regarding standards of practice for screening for fatty liver. The presence of fat in the liver can be suggested by various imaging modalities; however, no current noninvasive method can distinguish NASH from NAFLD. Liver biopsy remains the gold standard for staging and grading.

The presence, degree, and pattern of alanine aminotransferase (ALT) elevation are nonspecific. Even in those patients at high risk -- obese, diabetic individuals -- the ALT level does not distinguish between fatty liver alone and NASH. Hepatocellular glycogenosis causing hepatomegaly and abnormal liver tests can also be seen in poorly controlled diabetes mellitus and the ALT level does not discriminate hepatic glycogenosis from steatosis. Several studies presented during this year's DDW meeting reexamined the diagnostic sensitivity and specificity of liver enzyme levels. Kunde and colleagues[8] noted that of 272 morbidly obese subjects (237 females) undergoing Roux-en-Y gastric bypass, fatty liver was present in 33% and NASH in 36%. Fibrosis was observed in 64% of these subjects and 11% had advanced fibrosis (bridging/cirrhosis). The diagnostic sensitivity and specificity of ALT/aspartate aminotransferase (AST) for NASH was low. The spectrum of NAFLD was not significantly different in subjects with normal liver enzymes. In fact, within a cohort of 135 NAFLD patients studied by Pantsari and Harrison,[9] 10% presented with an isolated elevated alkaline phosphatase level. Harrison and colleagues[10] evaluated measures for accurately predicting the presence of advanced fibrosis in 135 patients with NAFLD. A composite index was created using biomarkers and clinical indices (age, AST/ALT, leptin and adiponectin levels, presence of diabetes, etc). This accurate noninvasive model for detecting fibrosis can be used to identify patients with NASH and clinically significant fibrosis.

Imaging
Transabdominal ultrasound is a sensitive, noninvasive method for detecting NAFLD. However, diagnostic criteria are highly operator-dependant and nonstandardized. Treiber and colleagues[11] quantified the amount of hepatic fat by abdominal ultrasound. Univariate analysis revealed a significant correlation between histologic degree of steatosis and ventral and dorsal liver areas; the predictive value to exclude a significant degree of steatosis was 98%. Using scintigraphy, Vetelainen and colleagues[12] reported a strong, significant inverse correlation between the severity of steatosis, hepatic triglyceride content, and 99mTc-mebrofenin uptake rate.

In the future, noninvasive "dynamic" breath tests may disclose specific alterations in metabolic pathways. For example, 13C-methacetin and 13C-KICA (ketoisocaproic acid) breath tests reflect microsomal and mitochondrial function, respectively. NASH patients have enhanced methacetin demethylation; conversely, KICA metabolism is significantly decreased in these patients.[13] Methacetin demethylation correlated positively with ALT levels and inversely with KICA values. The extent of KICA decarboxylation was inversely related to the severity of histologic damage. Ono and colleagues[14] used a branched-chain fatty acid analog as an imaging agent to assess mitochondrial function in 37 patients with NASH. The clearance rate, which was low as a result of impaired hepatic fatty acid beta-oxidation, may be used to identify NASH patients.

The Prevalence of NAFLD and NASH
Estimates based on autopsy studies suggest that 20% to 30% of adults have excess fat accumulation in the liver and 2% to 3% of adults meet diagnostic criteria for NASH.[15] In a prospective study of 241 consecutive patients in a primary care clinic, the prevalence of fatty liver was 37%.[16] Patients with NAFLD were more likely to be obese and have diabetes (54% vs 4%), hypertension (41% vs 10%), and insulin resistance (54% vs 20%). Multivariate analysis showed that diabetes, age, and waist circumference were independent predictors of NAFLD. In analysis of intraoperative biopsies obtained from 57 patients with body mass index (BMI) > 40 kg/m2 who underwent bariatric surgery, 11% showed no hepatocellular injury, 18% had steatosis, and 72% had NASH.[17] Diabetes, increased ALT levels, hyperlipidemia, and hypertension were more frequent in the group of patients with NASH.

What is the prevalence of NASH among different racial groups? Attar and colleagues[18] evaluated 1253 consecutive patients at Cook County Hospital (Chicago, Illinois) who presented with abnormal ALT levels -- 72 patients met criteria for NASH (53 female), 64% were Hispanic, 18% African American, 7% white, and 7% Asian. This distribution contrasts with their overall outpatient population -- 60% African American, 30% Hispanic, and < 10% white, Asian, or other. Significant fibrosis was associated with diabetes, higher BMI, and age > 45 years. However, Hispanics also had the highest risk for development of advanced fibrosis, and NASH occurred at a lower BMI and at younger age in Hispanics.

Low serum levels of adiponectin, an anti-inflammatory cytokine secreted by adipocytes, has been associated with progression of NAFLD. Baranova and colleagues[19] found that the adiponectin mRNA level in intra-abdominal adipose tissue of patients with diabetes was significantly lower than that in nondiabetic patients. Insulin resistance related to obesity and diabetes may be responsible for low levels of adiponectin, which in turn can potentially predispose to disease progression.

A lower prevalence of obesity-associated NAFLD is found in African-American children, compared with white and Hispanic children. Louthan and colleagues[20] reported that overweight African-American children had significantly lower adiponectin levels compared with white children, despite similar BMI Z score. In addition, the overweight white children had a significant negative correlation of adiponectin to fasting insulin and glucose; these correlations were not seen in African-American children. Better understanding of these differences may help explain prevalence differences in pediatric NAFLD between races.

The histologic features of NAFLD in children are different from those in older patients. A-Kader and colleagues,[21] in a review of 67 children (54 males) with NAFLD, noted that BMI did not correlate with ALT level or histologic grading. None of the children had ballooning degeneration, a common finding among adult patients with NASH. Necroinflammatory grade 1 was observed in 42% and grade 2 in 52% of subjects; fibrosis stage 0 was seen in 13% of subjects, stage 1 in 48%, and stage 2 in 26%. Intranuclear glycogen was observed in 21% of the biopsies. NAFLD was more common among the Hispanic population.

The Pathogenesis of NASH Is Multifactorial
Obesity and insulin resistance are key factors in exacerbating hepatic inflammation and fibrogenesis in NASH; insulin resistance is an antecedent in the accumulation of hepatocellular fat, whereas excess intracellular fatty acids, oxidant stress, adenosine triphosphate (energy) depletion, and mitochondrial dysfunction are important causes of hepatocellular injury in the steatotic liver. A "2-hit theory" is postulated: net retention of lipid in the hepatocyte triggers oxidative stress and cytokine release, culminating in NASH. There is also a role for genetic susceptibility and environmental influences.[15] NASH is associated with decreased insulin-mediated suppression of lipolysis; thus, subjects with NASH have high serum free fatty acid concentrations, allowing greater hepatic fatty acid uptake and oxidation. Increased fatty acid delivery to the liver affects the hepatocytes, interfering with insulin function and mitochondrial beta-oxidation. This leads to the development of steatohepatitis.

Morgan and colleagues[22] reported increased hepatic expression of transcription factors that activate hepatic genes involved in lipogenesis and fat accumulation in livers of mice with NAFLD. In another study, Feldstein and colleagues[23] determined the mechanism of fatty acid-induced steatosis -- lysosomal permeabilization leading to release of proteases into the cytosol. Specific regulatory genes were identified. Liver fatty acid binding protein (L-Fabp) is thought to facilitate the intracellular trafficking and metabolism of hydrophobic lipid molecules such as fatty acids. Newberry and colleagues[24] demonstrated that deletion of L-Fabp protects mice against the development of diet-induced hepatic steatosis; there was decreased hepatic expression of genes related to fatty acid oxidation. These investigators postulate that fatty acids are targeted for metabolism into cholesterol rather than for storage as triglyceride in the absence of L-Fabp. Pandya and colleagues[25] documented that under conditions of a high-fat diet, mice develop insulin resistance and hepatic injury analogous to human fatty liver disease due to upregulation of apolipoprotein C-III and downregulation of the p53 gene. These studies indicate potential therapeutic targets for lipotoxicity-mediated processes such as NAFLD.

What Genetic or Environmental Factors Play a Role in Progression From NAFLD to NASH?
Huang and colleagues[26] used a candidate gene approach to identify genetic markers associated with NASH in 187 subjects. DDX5, a helicase associated with fibrosis, was found to be associated with an increased risk of NASH, whereas MTP, a gene central to hepatic lipid export, was found to be associated with a decreased risk. Advanced glycation end products (AGEs) accumulate in diabetes and play an important role in the pathogenesis of angiopathy and insulin resistance. Additionally, AGEs activate hepatic stellate cells to induce fibrosis. In a study presented during these meeting proceedings, Hyogo and colleagues[27] reported that serum AGE levels are increased in patients with NASH. Hepatic AGE depositions were found in the pericellular areas and in the vicinity of portal areas; these findings were persistent regardless of stage and glucose intolerance patterns. AGEs may play an important role in the progression to NASH from NAFLD, and thus could be a predictive factor.

To further clarify the effect of AGEs in NASH, Iwamoto and colleagues[28] studied liver fibrosis and inflammation using hepatic stellate cells. AGEs stimulated proinflammatory/fibrogenic cytokines through overgeneration of reactive oxygen species; these effects were prevented by antioxidants. Trujillo and colleagues[29] found that the MnSOD (manganese superoxide dismutase) Ala16Val gene and TNF (tumor necrosis factor)-alpha gene G-polymorphisms were involved in determining susceptibility to NASH, likely inducing oxidative stress and inflammatory process. Thus, anti-AGE and antioxidative stress therapy may be useful in the treatment of NASH.

Increased iron stores correlate with the degree of fibrosis in NASH. Jayaraman and colleagues[30] hypothesized that higher iron stores in overweight diabetics would correlate with hepatic steatosis and higher levels of ALT. Of 131 subjects, 39% were iron-deficient, 31% anemic, and 15% both. For each 10% increase in transferrin saturation, mean ALT level increased by 12%. The mean ALT level in individuals with anemia was 18% lower than in those without anemia. Patients who had iron-deficiency anemia were less likely to have hepatic steatosis (37% vs 60%).

What Can We Learn From Models of Fatty Liver?
Steatosis is associated with familial partial lipodystrophy, which is caused by mutations in the LMNA gene. These patients have severe insulin resistance, dyslipidemia, and NASH. Peroxisome proliferator-activated receptor-gamma (PPAR-gamma) plays an important role in glucose metabolism and mutations within the PPARG gene cause familial partial lipodystrophy.[31] Thus, both LMNA and PPARG mutations can result in steatosis and may therefore serve as monogenetic models for steatosis. NASH is found in lean subjects without obvious metabolic disturbances. In a study presented during this year's DDW meeting, Friedrich and colleagues[32] found that altered body composition (decreased cell mass, as in lipodystrophy) to be a risk factor in 15 normal-weight patients (BMI < 27 kg/m2) with NASH.

Choline-deficient (CD) and methionine- and choline-deficient (MCD) diets induce 2 different models of steatosis; the CD diet causes slow progression of hepatic steatosis, whereas the MCD diet rapidly induces steatohepatitis.[33] The CD diet also induces key features of human metabolic syndrome -- obesity, insulin resistance, beta-cell dysfunction, and dyslipidemia. Okumura and colleagues[34] investigated the mechanism underlying the steatohepatitis caused by these diets. Hepatic steatosis and infiltration of leukocytes were prominent in mice fed an MCD diet; overt hepatic fibrosis was observed after 8 weeks. The mechanisms involved oxidative stress, TNF-alpha, and chemokines. Mice lacking physiologic upregulation of adiponectin levels lost body weight, suggesting that adiponectin plays a pivotal role in regulation of insulin sensitivity and modulation of inflammatory and profibrogenic responses in dietary steatohepatitis. The MCD diet increases hepatocyte expression of a Th-1 cytokine osteopontin (OPN), and OPN-knockout mice fed an MCD diet develop steatosis with significantly reduced hepatic injury and fibrosis.[35] OPN expression was found to stimulate hepatocyte injury but not steatosis in patients with NASH.

The Treatment of NAFLD
There is no effective treatment for NAFLD, but there are several potential approaches. They include: (1) treatment of risk factors -- weight reduction (diet and exercise, or in extreme cases, surgery); (2) pharmacologic treatment of insulin resistance; (3) use of lipid-lowering agents; and (4) use of drugs that protect hepatocytes/antioxidants (ursodeoxycholic acid [UDCA], betaine, vitamin E, lecithin, beta-carotene, and selenium). There were no significantly new data regarding the use of lipid-lowering agents in this setting presented during these meeting proceedings.

Weight Reduction
Improved liver histology occurs after weight reduction, but rapid weight loss can be associated with a transient worsening of inflammation and fibrosis. Andersson Friis-Liby and colleagues[36] studied the early changes in liver tests and in intrahepatic fat (by computed tomography) during rapid weight loss (overall weight loss was about 28 kg) in 40 patients with NAFLD. An initial increase of fatty infiltration in the liver was seen, in parallel to an increase in ALT levels. Thereafter, weight reduction induced normalization of liver fat and improved serum ALT and insulin sensitivity. Tendler and colleagues[37] prospectively enrolled 5 adult patients with biopsy-proven NASH to a very low carbohydrate diet (25 g/day). At 24 weeks, they observed improved ALT levels, lipid and insulin resistance parameters, hepatic steatosis, and histologic grade. This study could not discriminate between the effects of a low-carbohydrate diet and generalized weight loss.

In another study, Kaushik and colleagues[38] assessed the effects of Roux-en-Y gastric bypass surgery on liver histology in 31 obese patients with NAFLD. Mean BMI decreased from 51 to 34. All patients had steatosis on initial biopsy, but only 39% had steatosis on follow-up; 68% of subjects showed improvement in NASH grades and 23% had no inflammation following Roux-en-Y gastric bypass. Barker and colleagues[39] also reported that weight loss, achieved through Roux-en-Y gastric bypass, improved histopathology in 149 obese patients with biopsy-proven NASH. At the time of surgery, 23% of patients had histopathologic evidence of NASH. After an average of 642 days, histopathologic criteria for NASH were no longer found in 84% of patients. Surgery also improved hepatic steatosis and the resulting inflammation in 732 subjects evaluated by Keshishian.[40] No detrimental effects on hepatic function were noticed. Thus, in obese patients with NAFLD, gradual and substantial weight loss achieved by Roux-en-Y gastric bypass decreases hepatic fat content, inflammation, and fibrosis. A related study by Phillips and colleagues[41] aimed to examine the initial extent of postoperative fat loss from various anatomic sites. Fat loss was most marked from subcutaneous sites, but fat loss from visceral and hepatic depots was less marked. This finding presumably explains the lack of effect of this weight loss upon whole-body insulin sensitivity.

Pharmacologic Treatment of Insulin Resistance -- Improving Insulin Sensitivity Medically
The insulin sensitizer metformin* elicits resolution of hepatic steatosis in mice. In humans with NASH, metformin improves biochemical indices of hepatocellular injury and insulin resistance. Blaszyk and colleagues[42] treated 10 patients with biopsy-proven NASH with a 48-week course of metformin (2 g/day). Metformin improved hepatic necroinflammation but did not improve hepatic fibrosis. Thus, as sole therapy, metformin does not appear to be beneficial.

Thiazolidinediones (PPAR-gamma agonists) ameliorate insulin resistance and reduce hepatic fat stores in type-2 diabetes mellitus. Harrison and colleagues[43] performed a randomized, double-blind, placebo-controlled trial to examine the efficacy of a thiazolidinedione, pioglitazone* (45 mg for 6 months), in 22 patients with biopsy-proven NASH. Treatment with pioglitazone resulted in an approximately 2.5-fold increase in plasma adiponectin, reduced ALT levels, and a 25% reduction in hepatic fat content by magnetic resonance imaging. A significant improvement in ballooning degeneration, steatosis, and fibrosis was only seen with pioglitazone treatment. Discontinuation of pioglitazone in 21 patients was associated with a return of insulin resistance, increase in serum ALT levels, and a worsening of hepatic steatosis and inflammation.[44] Prolonged therapy with pioglitazone may be necessary to maintain the biochemical response and the histologic improvement seen with short-term therapy with this agent. Pentoxifylline* (1600 mg/day) reduced TNF-alpha levels and improved insulin sensitivity in patients, especially women, with biopsy-proven NASH.[45]

Drugs That Protect Hepatocytes/Antioxidants
Oxidative stress plays a major role in the pathogenesis of NASH; this results from free radicals generated by cytochrome P4502E1 (CYP2E1) induced by fatty acids. Lieber and colleagues[46] tested the hypothesis that combination s-adenosylmethionine and dilinoleoylphosphatidylcholine may be effective in treating NASH. Hepatic steatosis and CYP2E1 mRNA decreased with treatment, reflecting significant control of oxidative stress. Oliveira and colleagues[47] evaluated the efficacy of N-acetylcysteine* (1.2g/day orally) in reducing oxidative stress. They found that serum ALT levels were reduced in the first month of treatment, but were not decreased significantly at 6 months. Moreover, no improvement was noted in inflammation, fibrosis, or liver steatosis. Thus, N-acetylcysteine alone does not appear to be a potent therapy for NASH. Finally, Oz and colleagues[48] evaluated the potential efficacy of 2(RS)-n-propylthiazolidine-4®-carboxylic acid (PTCA),* an L-cysteine prodrug (an antioxidant involved in the biosynthesis of glutathione) in a dietary model of NASH. They found that PTCA therapy suppressed abnormal ALT levels to normal, improved pathologic findings, and reduced induction of genes involved in tissue remodeling.

Looking to the Future
The research agenda for NAFLD should focus on further defining the role of insulin resistance and abnormal lipoprotein metabolism in the pathogenesis of hepatocellular injury. The link between obesity and inflammation will continue to be defined.[49] Moreover, there will be good utility in understanding genetic factors and environmental influences predisposing to disease. Thus, insights provided by studies presented during this year's DDW meeting may provide the path forward for the development of noninvasive predictors of disease and more effective preventive and therapeutic strategies.