The association of hepatic steatosis with inflammatory changes and fibrosis in obese patients was first described over half a century ago. Although initially thought to be a relatively benign entity, reports began emerging that in some circumstances fat in the liver could lead to cirrhosis or liver failure, as in patients undergoing surgical jejunoileal bypass for morbid obesity.
The histologic features of “fatty liver disease” resemble alcohol-induced liver injury, but because they occur in patients with little or no alcohol consumption, the term nonalcoholic fatty liver disease (NAFLD) was coined. NAFLD is now recognized as one of the most common causes of chronic liver disease in the United States.
NAFLD encompasses a spectrum of diseases that ranges from bland hepatic steatosis, which is generally believed to be a benign condition, to hepatic steatosis with a necroinflammatory component that may or may not have associated fibrosis. This latter condition is termed nonalcoholic steatohepatitis (NASH) and is considered the ‘progressive’ form of NAFLD.
It is estimated that NAFLD affects up to 25% of the general population or 75 million Americans and that NASH affects approximately 8.6 million Americans, substantially higher than the estimated 1.6% prevalence of hepatitis C viral (HCV) infection. As a result, NASH cirrhosis is projected to surpass HCV as the leading indication for orthotopic liver transplantation (OLT) in the next 10 years.
The natural history and progression of NASH to cirrhosis is unclear with prevalence rates of 3% to 15% in case series and small prospective cohorts have shown that NASH may progress to cirrhosis in 9% to 20% of patients.
The pathogenesis of fatty liver disease and NASH is yet to be fully elucidated, but the common association with visceral obesity, hyperlipidemia, hypertension, and diabetes mellitus suggests that it is the hepatic manifestation of the metabolic syndrome, with at least 1 of these features present in over 90% of NAFLD patients.
Insulin resistance seems to play a pivotal role in the pathogenesis of NAFLD because it leads to hepatic steatosis. However, only a minority of such patients develops steatohepatitis and progressive liver disease; hence, other insults are required such as apoptosis, oxidative stress, and overexpression of inflammatory cytokines. Similarly, visceral obesity is implicated; it is associated with an increased release of a variety of adipokines [eg, adiponectin, leptin, tumor necrosis factor (TNF)-α], which seem to be important in the creating an environment for increased fibrosis.
Current therapy for NAFLD includes weight loss through diet and exercise and, in more extreme cases, bariatric surgery. Because there is such a strong association between NAFLD and the metabolic syndrome, many treatment options under investigation are aimed at improving insulin sensitivity; other studies have examined the effect of anti-inflammatory or anti-oxidant agents. This article reviews the current literature on pharmacologic and surgical options for NAFLD.
Pharmacologic Treatments
The need for treatment in NAFLD is based on the concern for progressive liver disease and cirrhosis. Natural history studies indicate this occurs in a minority of patients, but the high prevalence of the disease means an effective treatment could have major economic and health benefits.
Several agents with differing mechanisms of action have been investigated in NAFLD and reflect the different pathways that are thought to be important in pathogenesis. Improving insulin resistance and reducing oxidative stress are the 2 main avenues that have been studied, but several other drugs have been examined.
Thiazolidinediones
Thiazolidinediones (TZDs) function as selective agonists for peroxisome proliferator activated nuclear receptor-γ. Various pathways have been implicated in hypothesizing TZDs actions in NAFLD, but the overriding actions seem to be decreasing hepatic fatty acid levels (by decreasing lipolysis and increasing β-oxidation), redistributing fat content from the liver to peripheral adipose tissue, and promoting insulin sensitivity by facilitating preadipocyte differentiation into insulin-sensitive adipocytes. TZDs have also been shown to enhance adiponectin levels, a cytokine whose concentration is decreased in NASH patients and may play a role in decreasing hepatic fat content. Pioglitazone and rosiglitazone have been extensively studied in NAFLD and the trials are summarized in Tables 1 and 2 . Unfortunately, the majority of trials lacked power and lacked placebo controls. Belfort and colleagues conducted a placebo-controlled trial where they looked at diabetic patients with NASH proven on entry biopsy. Six months of pioglitazone resulted in a significant decrease in steatosis, ballooning injury, and centrilobular inflammation as seen on repeat biopsy. Pioglitazone treatment also significantly improved liver transaminases, markers of insulin sensitivity, and adiponectin levels. However, there was no difference in fibrosis compared with placebo and pioglitazone treatment significantly increased body fat. In another study looking at nondiabetic patients with NASH, 1 year of pioglitazone treatment produced histologic improvement in ballooning injury, Mallory bodies, and fibrosis, but was no better than placebo in terms of steatosis and inflammation. Transaminase levels were once again significantly reduced, but there was no difference in markers of insulin sensitivity; adiponectin levels were increased, but the change was not significant compared with placebo. Significant weight gain was again seen in the TZD group. In the PIVENS trial, 2 years of pioglitazone therapy in nondiabetic NASH patients resulted in significant improvement in liver transaminases, steatosis, inflammation and overall NAFLD activity score (NAS). Pioglitazone use was also associated with significant resolution of NASH compared with vitamin E or placebo. Ballooning scores improved with pioglitazone but did not reach significance and there was no improvement in fibrosis. Again, significant weight gain was seen despite an improvement in insulin resistance.
| Author | Year | Design | Patients | Duration (mos) | Liver Enzymes | Steatosis | Necroinflammation | Fibrosis |
|---|---|---|---|---|---|---|---|---|
| Sanyal et al | 2004 | RCT | 20 | 6 | − | + | + | + |
| Belfort et al | 2006 | RCT | 55 | 6 | + | + | + | − |
| Aithal et al | 2008 | RCT | 61 | 12 | + | − | +/− | + |
| Sanyal et al | 2010 | RCT | 80 | 24 | + | + | + | − |
| Author | Year | Design | Patients | Duration (mos) | Liver Enzymes | Steatosis | Necroinflammation | Fibrosis |
|---|---|---|---|---|---|---|---|---|
| Neuschwander-Tetri et al | 2003 | Open | 30 | 12 | + | n/a | + | + |
| Ratziu et al | 2008 | RCT | 63 | 12 | + | + | − | − |
| Ratziu et al | 2010 | RCT | 53 | 24–36 | + | + | − | − |
| Omer et al | 2010 | RCT | 64 | 12 | + | + | n/a | − |
Rosiglitazone has been studied in the FLIRT trials. In the original study, 1 year of rosiglitazone treatment resulted in significant improvement in steatosis, transaminase levels, adiponectin levels, and insulin sensitivity. However, there was no improvement in histology based on overall NAS. There was an average weight gain of 1.5 kg in the rosiglitazone arm. An open-label extension of this trial for 2 additional years maintained the significant reductions in steatosis and transaminase levels and, although the NAS improved over the 2 years, the trend was not significant. Among patients who had a greater than 30% reduction in steatosis, significant improvement was shown in necroinflammation and a nonsignificant improvement was seen in ballooning. Weight gain persisted over the 3-year period. In a recent randomized trial comparing rosiglitazone with metformin, rosiglitazone therapy for 1 year significantly reduced transaminase levels, insulin resistance and overall NAS without any change in body mass index (BMI).
Although its effects on ballooning, fibrosis, and adiponectin levels remain controversial, TZD therapy seems to improve insulin sensitivity, liver enzymes, steatosis, and necroinflammation, with pioglitazone being consistently more effective than rosiglitazone.
Discontinuation of TZDs results in a return to pretreatment NASH histology and serum markers, suggesting that TZD therapy would have to be maintained indefinitely for continued treatment response. This is likely to result in weight gain, which is due to increased total body fat and not fluid retention. Because increased total body fat and insulin resistance are the initial triggers in the NAFLD cascade, it is unclear how TZD-induced weight gain would affect NAFLD patients long term and whether the histologic benefits would eventually be reversed. Concerns over cardiovascular events and bone pathology also surround TZD use, but have not been adequately studied in NAFLD patients.
TZD therapy, especially with pioglitazone, seems to be an effective treatment for NASH but needs to be given indefinitely and, given its propensity for weight gain, it needs further study. It can be considered one of the preferred agents in diabetic patients with NASH because these patients already have an indication for TZD therapy.
Vitamin E
Vitamin E (α-tocopherol) is a naturally occurring antioxidant. Its effects in NASH may be secondary to its function as a free radical scavenger or its ability to inhibit cytokines such as transforming growth factor (TGF-β), which plays a role in hepatic stellate cell activity and fibrogenesis as shown in rat models. NASH patients have also been shown to have decreased levels of vitamin E and increased levels of TGF-β as compared with healthy controls, and 1 year of vitamin E treatment significantly reduced TGF-β levels in NASH patients. Vitamin E may also have a non-antioxidant mechanism in NASH; it may alter insulin resistance by influencing proliferator activated nuclear receptor-α expression.
Of the multiple open-label and randomized controlled trials ( Table 3 ), the largest trial looking at the effects of vitamin E in patients with NASH was the PIVENS trial. A total of 247 nondiabetic patients with NASH were randomized to receive vitamin E, pioglitazone, or placebo daily for 2 years. Vitamin E at 800 U/d resulted in a significant reduction in liver enzymes, steatosis, ballooning, lobular inflammation, and overall NAS. Although none of the 3 arms had any effect on fibrosis, only vitamin E resulted in a significant improvement in ballooning scores and there was no side effect of weight gain as seen in the pioglitazone arm. Although PIVENS showed no effect on fibrosis, Harrison and colleagues showed that vitamin E at 1000 U/d along with vitamin C 1 g/d significantly improved fibrosis at 6 months. In this trial, which included diabetic patients, vitamin E therapy resulted in significant improvement in fibrosis scores without any changes in inflammation, necrosis, or transaminase levels. Interestingly, the effect on fibrosis was most profound in diabetic patients. Steatosis was not evaluated in the study and minimal improvement in fibrosis scores was also seen in the placebo group. Another open label trial of 300 U/d of vitamin E for 1 year also showed improvement in transaminase levels, steatosis, inflammation, and fibrosis, but was limited by small sample size and lack of paired biopsies.
| Author | Year | Design | Patients | Duration (mos) | Liver Enzymes | Steatosis | Necroinflammation | Fibrosis |
|---|---|---|---|---|---|---|---|---|
| Hasegawa et al | 2001 | Open | 12 | 12 | + | + | + | + |
| Harrison et al | 2003 | RCT | 49 | 6 | − | n/a | − | + |
| Sanyal et al | 2004 | RCT | 20 | 12 | n/a | + | − | − |
| Bugianesi et al | 2005 | RCT | 28 | 12 | + | n/a | n/a | n/a |
| Sanyal et al | 2010 | RCT | 84 | 24 | + | + | + | − |
Vitamin E has also been studied in combination with other NAFLD agents with mixed results. A randomized, controlled trial with ursodeoxycholic acid (UDCA) and vitamin E demonstrated that the combination was more effective than UDCA alone in terms of steatosis and transaminase reduction but there was no significant difference in ballooning, inflammation, or fibrosis. A pilot study looking at vitamin E and pioglitazone versus vitamin E alone in nondiabetic patients showed that the combination was superior to vitamin E alone in terms of NASH histology. Patients receiving both vitamin E and pioglitazone had significant improvements in steatosis, ballooning, inflammation, and fibrosis scores, whereas those on vitamin E alone only significantly reduced their steatosis grade. Not surprisingly, the combination with pioglitazone also demonstrated improved insulin sensitivity compared with vitamin E alone.
Vitamin E treatment seems to be beneficial in NASH, but there is a note of caution. High-dose vitamin E therapy has been associated with increased mortality in other studies, and most NAFLD trials have used doses of vitamin E above the current recommended dose. It is reassuring that the PIVENS trial did not report any increase in mortality from vitamin E use, but may have been underpowered for this end point. It would be prudent until more data emerges to use vitamin E selectively in NASH patients with more severe changes on histology. Combination regimens including vitamin E cannot currently be recommended.
Metformin
Metformin, a biguanide, is traditionally considered first-line treatment for non–insulin-dependent diabetes. Because there is a high prevalence of diabetes in patients with NAFLD, targeting insulin resistance with metformin seems like an appropriate pharmacologic option. Metformin increases the activity of 5′AMP-activated protein kinase, resulting in decreased hepatic glucose production and hepatic insulin resistance. This improvement in insulin sensitivity decreases hepatic lipid accumulation and, therefore, liver steatosis. Moreover, without the steatosis, hepatocytes are theoretically less susceptible to reactive oxygenated species, TGF-β, and other endotoxins that trigger the NASH cascade.
Many trials assessing metformin therapy in NAFLD have been conducted ( Table 4 ), but again are hampered by small sample size and poor study design. The first trial using metformin in humans with NAFLD was a small, open-label study that showed that metformin treatment resulted in a significant reduction in transaminase levels, insulin resistance, and liver volume seen on ultrasound compared with a diet-only control group. Another study comparing metformin with diet alone examined histologic end points in a small sample of patients. At the end of 6 months, the metformin group had significant reductions in transaminase levels, insulin resistance, and steatosis as seen on ultrasonography. In terms of histology, however, there was no difference between the metformin and diet groups. More patients actually had worsening fibrosis scores in the metformin group compared with the diet-only group, although this trend was not significant. Another open-label trial looked at metformin therapy for 1 year and had a small number of post-treatment liver biopsies. Metformin treatment reduced steatosis in one third, inflammation in one fifth, and fibrosis in just 1 patient. The effect on transaminase levels was mild with an initial decrease and then a return to pretreatment levels by the end of the study. Loomba and colleagues demonstrated a more profound histologic effect in their open-label study. One year of metformin treatment improved the histology in more than 50% of patients and one third of patients showed an impressive 3-point decrease in the NASH activity index. Fibrosis scores decreased minimally, but the changes were not significant. Although insulin sensitivity significantly improved, transaminase levels did not change over the year. However, those patients who histologically responded to metformin had significantly lower transaminase levels as well as a much greater improvement in insulin sensitivity. In another controlled trial, 55 patients were randomized to receive metformin, whereas others received vitamin E or a strict diet regimen. Metformin resulted in significant improvements in insulin sensitivity and the most rapid reduction in transaminase levels. However, at the end of 1 year, the decreased transaminase levels in the metformin group were not significantly different than those reduced levels seen in the diet group. Metformin treatment significantly reduced steatosis, inflammation, and fibrosis, but a very limited number of repeat biopsies were performed.
| Author | Year | Design | Patients | Duration (mos) | Liver Enzymes | Steatosis | Necroinflammation | Fibrosis |
|---|---|---|---|---|---|---|---|---|
| Marchesini et al | 2001 | Open | 20 | 4 | + | n/a | n/a | n/a |
| Uygun et al | 2004 | RCT | 36 | 6 | + | + | n/a | n/a |
| Nair et al | 2004 | Open | 15 | 12 | − | + | − | − |
| Bugianesi et al | 2005 | RCT | 55 | 12 | + | + | + | + |
| Loomba et al | 2008 | Open | 28 | 12 | − | +/− | + | − |
| De Oliveira et al | 2008 | Open | 20 | 12 | + | + | − | + |
| Haukeland et al | 2009 | RCT | 48 | 6 | − | − | − | − |
| Garinis et al | 2010 | RCT | 50 | 6 | n/a | + | n/a | n/a |
| Omer et al | 2010 | RCT | 22 | 12 | − | − | − | − |
Recent trials continue to provide mixed results. Omer and colleagues compared metformin with rosiglitazone in a 1-year randomized trial of diabetic NASH patients. Metformin significantly reduced the waist circumference and BMI of patients, but had no effect on transaminase levels or NAS. On the other hand, the combination of metformin with rosiglitazone significantly improved transaminase levels and NAS, suggesting that metformin may have a role in potentiating TZD effects in NAFLD. In another study looking at ultrasound-based assessment of steatosis, metformin was no better than diet in terms of improving steatosis, although it did significantly improve insulin resistance and increase adiponectin levels.
Metformin may improve insulin sensitivity in NAFLD, but its effects on transaminases, steatosis, inflammation, and fibrosis remain unclear. Further trials of longer duration and larger sample sizes that look at histologic outcomes are necessary before metformin can be recommended for use in NAFLD.
Probucol
Probucol is a lipid-lowering agent with potent antioxidant properties, allowing it to function as a free radical scavenger. Moreover, its affinity to accumulate in adipose tissue makes it an attractive candidate for use in NAFLD.
Merat and colleagues have conducted 3 trials of probucol in NAFLD. In their original pilot study, 500 mg of probucol for 6 months significantly reduced transaminase levels as well as total cholesterol. However, most of the reduction in cholesterol was due to a drop in high-density lipoprotein cholesterol (HDL); low-density lipoprotein cholesterol and triglyceride levels did not change significantly. In a subsequent placebo-controlled trial, probucol significantly reduced alanine aminotransferase levels, but there were no changes in aspartate aminotransferase or total cholesterol levels when compared with placebo. HDL levels did decrease in the treatment arm, but the changes were not significant. In the only trial that looked at liver histology, 8 patients received probucol for 1 year and then underwent repeat biopsies. The overall necroinflammatory grade, based on a modified Brunt scoring system that looked at steatosis, ballooning, and inflammation, was significantly improved with probucol treatment. Transaminase levels and steatosis as assessed by ultrasonography were also significantly reduced. There were no changes in the stage of fibrosis, total cholesterol, or HDL levels, but low-density lipoprotein cholesterol levels did decrease significantly. No significant side effects were seen.
Although the trials have been limited by small sample sizes and short durations, they do consistently show a reduction in transaminase levels and the 1 study that evaluated liver histology also demonstrated favorable effects. However, there are concerns over the use of probucol. It is currently not available in many countries, including the United States, because there are more efficacious lipid-lowering alternatives and because of its potential adverse effects. Probucol can lower HDL levels and cardiac arrhythmias have also been seen. Therefore, until randomized, controlled trials show consistent histologic benefits and no adverse effects, probucol cannot be recommended for use in NAFLD.
Pentoxifylline
Pentoxifylline is a phosphodiesterase inhibitor, resulting in decreased activity of TNF-α and TGF-β. TNF-α is a potent inflammatory cytokine that is upregulated in NAFLD and is a mediator of insulin resistance. TGF-β, a primary target of vitamin E therapy in NAFLD, seems to play a role in hepatic stellate cell activity and fibrogenesis. Therefore, pentoxifylline could theoretically function as an antioxidant, antifibrotic, and insulin-sensitizing agent in NAFLD.
The majority of trials evaluating the effects of pentoxifylline in NAFLD have been limited by small sample size, short duration, lack of a placebo control, and no analysis of histologic outcomes ( Table 5 ). One of the original pilot studies used 1600 mg/d of pentoxifylline for 1 year and showed a significant reduction in transaminase levels, but one third of the patients withdrew from the study because of nausea. No changes were seen in fasting glucose, triglyceride, or HDL cholesterol levels. Using 1200 mg/d also significantly decreased transaminase levels and the reduced dose did not cause nausea. Fasting serum insulin, the homeostasis model assessment of insulin resistance (HOMA-IR), and TNF-α levels were also significantly reduced by pentoxifylline in this trial. An extension of this trial was then conducted to look at the histologic effects of pentoxifylline. Of the original 15 patients that received 1200 mg/d for 6 months, 9 continued treatment for a full year and underwent repeat liver biopsies. A majority showed a reduction in steatosis, necroinflammation, and fibrosis. Transaminase levels and insulin sensitivity also significantly improved. There were no significant side effects seen at the 1200 mg/d dose and no patients dropped out of the study. In one of the few placebo-controlled trials, pentoxifylline significantly reduced transaminase levels, but had no effect on TNF-α, IL-6, or hyaluronic acid levels when compared with placebo.
| Author | Year | Design | Size | Duration (mos) | Transaminases | Steatosis | Necroinflammation | Fibrosis |
|---|---|---|---|---|---|---|---|---|
| Adams et al | 2004 | Open | 20 | 12 | + | n/a | n/a | n/a |
| Satapathy et al | 2004 | Open | 18 | 6 | + | n/a | n/a | n/a |
| Satapathy et al | 2007 | Open | 9 | 12 | + | + | + | + |
| Lee et al | 2008 | RCT | 20 | 3 | + | n/a | n/a | n/a |
Related posts:
Emerging Therapies in Hepatitis C: Dawn of the Era of the Direct-Acting Antivirals
Noninvasive Tools to Assess Hepatic Fibrosis: Ready for Prime Time?
Liver Transplantation in the 21st Century: Expanding the Donor Options
Hepatitis B: Modern End Points of Treatment and the Specter of Viral Resistance
Liver Transplantation in the 21st Century: Expanding the Donor Options
Alcoholic Hepatitis: Prognostic Models and Treatment
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
