- 1H MRS
proton magnetic resonance spectroscopy
- IR
insulin resistance
- NAFLD
non-alcoholic fatty liver disease
- NASH
non-alcoholic steatohepatitis
The mounting clinical and research interest in fatty liver disease is a consequence of the worldwide epidemic of obesity. Fatty liver is one of the most common liver disorders in the western world. In the absence of specific causes of hepatic fat accumulation such as high alcohol consumption, this condition is termed non-alcoholic fatty liver disease (NAFLD).
NAFLD represents a spectrum of liver disease of increasing severity that encompasses steatosis (fatty change), non-alcoholic steatohepatitis (NASH) with or without fibrosis and cirrhosis(Reference Angulo1). Most frequently, NAFLD is associated with obesity, insulin resistance (IR), dyslipidaemia and hypertension and should be considered the hepatic manifestation of the metabolic syndrome(Reference Angulo1). Alternative aetiologies of fatty liver include high alcohol consumption, rapid weight loss, total parenteral nutrition and drugs (Table 1)(Reference Sanyal2).
Table 1. Aetiology of non-alcoholic fatty liver disease and non alcoholic steatohepatititsFootnote *
HAART, highly active antiretroviral therapy.
* Table includes data from reviews(Reference Angulo1, Reference Clark, Brancati and Diehl3).
Although highly prevalent, NAFLD/NASH is frequently asymptomatic. Until recently it was considered to be an innocent bystander in the multi-system metabolic syndrome with few clinical sequelae. It is now generally accepted that NAFLD is not a benign condition and poses a major challenge to health care provision worldwide. This review will consider the importance of NAFLD/NASH in terms of (1) the prevalence and natural history; (2) associated morbidity and mortality and (3) potential for successful treatment, with an emphasis on nutritional factors.
How prevalent is fatty liver?
Estimates of prevalence rely on both an accurate case definition (i.e. how a case of fatty liver is defined) and sample frame (i.e. the population from which estimates of prevalence are derived).
Case definition
There is no universally applicable diagnostic test for hepatic steatosis. A number of techniques with varying degrees of accuracy and applicability have been used. Generally, those tests that supply most information are less acceptable to patients (such as liver biopsy) or prohibitively expensive for population-wide screening (for example, proton magnetic resonance spectroscopy (1H MRS)), while acceptable tests such as waist-to-hip measurements or simple blood variables may lack precision and/or accuracy. Elevated liver enzymes in the absence of other causes may reveal only a subset of individuals with hepatic steatosis. The choice of test will therefore influence the estimate of prevalence.
Histology
Histological assessment is considered the reference standard for the assessment of NAFLD. Hepatic steatosis is defined pathologically as the presence of large and small vesicles of fat, predominantly TAG, accumulating within hepatocytes(Reference Clark, Brancati and Diehl3). The National Institute of Diabetes and Digestive and Kidney Diseases NASH Research Network has developed a well-validated scoring system that facilitates uniform and reproducible diagnosis of NAFLD/NASH using H&E (haematoxylin and eosin) stained tissue(Reference Kleiner, Brunt and Van Natta4). This score measures disease severity and quantifies fat in terms of the percentage of hepatocytes containing visible lipid droplets, together with features indicative of inflammation (presence of inflammatory cells and hepatocyte ballooning degeneration) and fibrosis(Reference Kleiner, Brunt and Van Natta4). Recently, the addition of digital image analysis techniques using Oil Red O stains to improve accuracy of histological steatosis quantification, at least within clinical trials, has been suggested(Reference Levene, Kudo and Thursz5). Even so, as liver biopsy typically samples only 1/50 000th of the liver mass and fat is not always evenly distributed, sampling error may occur. Further, liver biopsy is an invasive technique and carries a small but defined morbidity and so is inappropriate for population screening(Reference Gilmore, Burroughs and Murray-Lyon6).
Ultrasound
Ultrasound is widely available, acceptable to patients and offers the prospect of providing additional morphological information. However, the assessment of the degree of steatosis relies on observing comparatively increased echogenicity of the fatty liver compared with the renal cortex. The assessment is therefore subjective and semi-quantitative at best(Reference Schwenzer, Springer and Schraml7, Reference Mehta, Thomas and Bell8). The sensitivity and specificity of ultrasound is the subject of some controversy, but sensitivity for significant steatosis (equivalent to >30% on biopsy) is thought to be between 60% and 90%(Reference Schwenzer, Springer and Schraml7).
Magnetic resonance spectroscopy and imaging
1H MRS and MRI offer quantitative assessment of hepatic TAG(Reference Schwenzer, Springer and Schraml7, Reference O'Regan, Callaghan and Wylezinska-Arridge9, Reference Thomas, Hamilton and Patel10) and potentially lipid composition(Reference Cobbold, Patel and Goldin11). The techniques are relatively expensive, but non-invasive, and 1H MRS has been used in population-based studies(Reference Browning, Szczepaniak and Dobbins12, Reference Szczepaniak, Nurenberg and Leonard13). 1H MRS provides a measurement of TAG and water in a region of interest, which are influenced by the tissue assessed and the sequences used. Multi-echo MRI techniques also measure fat content comparing the signal phase and relaxation time from each component (fat and water) to calculate the relative abundance(Reference Guiu, Petit and Loffroy14). This approach has the benefit of enabling fat distribution to be visualised and reduce the need for post-processing. For absolute quantification, calibration to a known standard is required.
Biometrics and blood test indices
Routine measurements such as waist circumference, waist:hip ratio, BMI, together with simple blood-derived indices such as serum TAG levels, serum gamma glutamyl transpeptidase, and raised serum alanine aminotransferase and aspartate aminotransferase enzymes measurements, without other explanation, may also be used to predict NAFLD(Reference Bedogni, Bellentani and Miglioli15, Reference Church, Kuk and Ross16). Such measures are surrogate markers of hepatic steatosis and are not considered quantitative.
Furthermore, biometric and blood test indices have been applied to discriminate between patients with NAFLD and mild or no fibrosis from those with severe fibrosis, at greater risk of progression to cirrhosis and hepatocellular carcinoma. The NAFLD Fibrosis Score is an algorithm based on age, BMI, presence of hyperglycaemia, aspartate aminotransferase, alanine aminotransferase, platelet count and albumin. The model, built using a cohort of 480 patients with biopsy-proven NAFLD and validated in an independent cohort of 253, had an overall diagnostic accuracy of 0·88 and 0·82, respectively, in the two cohorts(Reference Angulo, Hui and Marchesini17). While this marker panel uses widely available clinical data, other markers aim to measure the intermediates of fibrogenesis, such as the Enhanced Liver Fibrosis panel, validated for the detection of moderate to severe fibrosis in NAFLD(Reference Guha, Parkes and Roderick18). Scoring systems such as these are being used as surrogate markers in clinical studies and may reduce the requirement for liver biopsy in a subset of patients.
Sample frame
The sample frame must be considered while estimating the prevalence of NAFLD in the general population. Differences in race, geography and socio-economic grouping may contribute different risk factors and thus they may impact on the results of population-based assessments of hepatic steatosis, while the prevalence of hepatic steatosis in at risk groups such as those with type-2 diabetes and obesity, is higher than the general population, making extrapolation difficult. This variation has similarly been recognised within the diagnostic criteria for the wider metabolic syndrome, reflected in the range of diagnostic criteria that have been produced. Thus, the International Diabetes Foundation (Table 2) criteria have been established as a single internationally adopted definition for future studies(Reference Alberti, Zimmet and Shaw19).
Table 2. Diagnostic criteria for the metabolic syndromeFootnote *
WHO, World Health Organization; NCEP, National Cholesterol Education Program; IDF, International Diabetes Foundation; OGTT, oral glucose tolerance test. M, male; F, female.
* Table modified from Bugianesi et al. 2005(Reference Bugianesi, McCullough and Marchesini72).
Prevalence assessed by population
Angulo estimated the prevalence of NAFLD in the US population to be about 30 million (or 21% of adults) in 2000 on the basis of the prevalence of obesity in the US population, multiplied by the proportion of those found to have hepatic steatosis in studies of obese people(Reference Angulo1). Clark and coworkers. estimated the prevalence to be 31 million (or 23% of adults) in the US on the basis that 79% of patients with mildly abnormal liver function tests and no other cause described were likely to have NAFLD(Reference Clark, Brancati and Diehl3). In at-risk groups, the estimates of prevalence are markedly higher with obese and morbidly obese subjects having a prevalence of 60–90%, respectively, while half of those with type-2 diabetes have steatosis. In obese diabetic patients, the presence of hepatic steatosis is universal(Reference Angulo1).
Analysis performed as part of the Dallas Heart Study examined 2349 subjects. Of these, 345 had no risk factors for fatty liver. Hepatic TAG content was measured by 1H MRS and a threshold, quoted as 5·56% TAG, found at the 95th centile, used to define abnormality. When applied back to the larger population, 34% of the individuals were found to have hepatic fat in excess of this level and thus hepatic steatosis(Reference Szczepaniak, Nurenberg and Leonard13). Still using 1H MRS, the prevalence of hepatic steatosis was studied in different races: prevalence in black people was 24%, 33% in the white population (42% in males and 24% in females), while in the Hispanic group, the prevalence was estimated at 45%, regardless of gender(Reference Browning, Szczepaniak and Dobbins12). The Dionysos study in northern Italy estimated a prevalence of 20–25% using ultrasound in a population of over 3000 individuals with and without suspected liver disease. It is estimated that between 43% and 55% of NAFLD patients with raised aminotransferases have histological evidence of NASH, while necroinflammation on histology from initial liver biopsy is the strongest predictor of fibrosis progression(Reference Ratziu, Bellentani and Cortez-Pinto20).
Pathogenesis
When discussing whether fatty liver is important, the pathogenesis of the condition should be considered to determine whether fat per se is pathogenic in progressive steatohepatitis, or whether it is the metabolism, utilisation and indeed the type of fat that contributes to disease. A foundation for research in the field has been the ‘Two-Hit Hypothesis’ proposed by Day and James(Reference Day and James21), where an initiating event promotes fat deposition which in turn sensitises the liver to hepatocellular damage and inflammation. Although our understanding of the complexity of NAFLD/NASH pathogenesis has expanded, there has been a move away from the linear model of pathogenesis; it remains a useful schema for clinical practice. A full discussion of the pathogenesis of NAFLD/NASH is outside the scope of this article, but has recently been comprehensively reviewed(Reference Fabbrini, Sullivan and Klein22), while key pathways are summarised in Fig. 1(Reference Cheung and Sanyal23).
Fig. 1. Key pathways in the pathogenesis of non-alcoholic fatty liver disease (NAFLD)/non-alcoholic steatohepatitis (NASH). Development of insulin resistance (IR) through genetic and environmental means, leads to increased de novo lipogenesis and TAG storage, with decreased oxidation and export, leading to net lipid accumulation. Increases in NEFA lead to endoplasmic reticulum (ER) stress and the production of reactive oxygen species, which cause hepatocellular inflammation and apoptosis. Pro-inflammatory cytokines, such as TNFα via NF-κB, further contribute to IR. These factors also lead to the activation of hepatic stellate cells (HSC) and net deposition of pathological extracellular matrix (fibrosis). Beneficial therapies act at multiple points, including those indicated on the right hand side of the Figure, leading to decreased IR and further beneficial downstream effects. Adapted from Cheung and Sanyal(Reference Cheung and Sanyal23).
NAFLD/NASH is best considered a complex disease trait where host genetic factors and the environment interact to produce disease phenotype and is therefore subject to subtle inter-patient variations beyond those that may be addressed by lifestyle modification(Reference Anstee and Goldin24). Central to the pathogenesis of NAFLD/NASH are the combined effects of IR, both hepatic and peripheral and intracellular NEFA accumulation. In general, the development of hepatic steatosis occurs when the rate of synthesis or import of NEFA by hepatocytes exceeds the rate of export or catabolism(Reference Anstee and Goldin24). In the presence of IR, insulin-mediated suppression of hormone sensitive lipase falls, increasing the release of NEFA from adipose stores into the circulation and subsequent uptake into hepatocytes(Reference Day25). Hyperinsulinaemia generated by compensatory pancreatic beta-cell hyper-secretion, mediates increased de novo lipogenesis through the transcription factor, sterol regulatory element-binding protein-1c and once beta-cell failure develops, hyperglycaemia acts via carbohydrate response element-binding protein to further promote lipogenesis. In parallel, mitochondrial fatty acid β-oxidation oxidative capacity becomes overwhelmed and alternative peroxisomal β-oxidation and microsomal ω-oxidation pathways assume greater importance, both of which lead to the increased production of reactive oxygen species(Reference Reddy26).
Does the presence of hepatic steatosis lead to inflammation?
The progression from simple steatosis to steatohepatitis is not simply due to the accumulation of NEFA-derived TAG within hepatocytes (whether derived from de novo fatty acid synthesis or flux of NEFA from peripheral adipose tissue), nor does it require a distinct second ‘hit’. Visible TAG deposition within hepatocytes may be considered an innocent bystander. This was elegantly demonstrated by small-interfering RNA interference of diacylglycerol transferase 2 expression, a key enzyme in TAG synthesis, in genetically obese/diabetic leptin receptor deficient Db/Db mice fed a methionine-choline deficient diet to induce steatohepatitis(Reference Yamaguchi, Yang and McCall27). Although steatosis was ameliorated, histological and biochemical markers of tissue oxidative stress, hepatocellular damage and apoptosis were increased. Hepatocellular damage and inflammation are a consequence of the accumulation of NEFA and lipid peroxidation products (lipotoxicity), leading to activation of inflammatory pathways, mitochondrial dysfunction, endoplasmic reticulum stress and ultimately apoptosis(Reference Cusi28). NEFA are produced as part of normal hepatic glucose metabolism via acetyl CoA and the citric acid cycle. NEFA are produced in excess in obesity, systemic inflammation, insulin resistant states and certain diets (such as high fructose diet). NEFA are potentially hepatotoxic(Reference Leclercq, Farrell and Field29), so experiments such as this suggest that the production and packaging of relatively inert TAG in the form of intracellular lipid droplets represents an adaptive response to the presence of NEFA. One intriguing hypothesis is that maternal nutritional factors may affect disease progression in adult offspring, possibly by the ‘priming’ of maternally inherited mitochondria. A study demonstrated that among mice fed a high fat diet, offspring of female mice fed a high fat diet went on to develop steatohepatitis, while offspring of those fed a control diet developed simple steatosis only. This was supported by the observation that mitochondrial electron transport chain enzymes were decreased in offspring of females fed high fat diet, while expression of genes involved in oxidative stress and inflammation were elevated(Reference Bruce, Cagampang and Argenton30). Case–control studies in the first instance would be required to demonstrate such an association in human subjects.
What type of fat is important?
NEFA are considered to influence progression of disease, but recent studies have established difference in the lipid profiles of patients with differing disease severity, suggesting that fat type may also be important. Murine and human studies have both demonstrated lower levels of hepatic PUFA with increasing steatosis and further, with steatohepatitis, compared to healthy controls(Reference Cobbold, Anstee and Goldin31, Reference Allard, Aghdassi and Mohammed32). Lipid profiling of human plasma showed that increased lipogenesis, desaturases and lipooxygenase activities characterised NAFLD and NASH, while impaired peroxisomal PUFA metabolism and non-enzymatic oxidation was associated with progression to NASH(Reference Puri, Wiest and Cheung33). These cross-sectional studies suggest that interventions to alter hepatic lipid composition and quantity may alter the natural history of the condition.
Does hepatic steatosis increase morbidity and mortality?
Most epidemiological studies of NAFLD are cross-sectional rather than prospective longitudinal. Accordingly, while the prevalence of NAFLD-associated conditions may be estimated subject to the considerations of case definition and sample frame, the proportion of patients progressing to more severe disease is not known. In an example of selection bias, patients with biopsy-proven NAFLD have worse outcomes than those who do not undergo biopsy, reflecting the selection criteria for liver biopsy. Two longitudinal studies by Adams and coworkers. warrant further description. In a population-based cohort study, 345 subjects with NAFLD diagnosed between 1980 and 2000 were followed up for a mean of 7·6 years. All-cause mortality was increased compared to the general population, while in patients with NAFLD, liver-related mortality was the third biggest cause of death, compared to the thirteenth in the general population(Reference Adams, Lymp and St Sauver34). In a study of over a hundred patients with paired biopsies an average of 3·2 years apart, progression of fibrosis was seen in 37%, regression in 29% and the remainder were unchanged. Risk factors for progression were low initial fibrosis stage, diabetes and increased BMI(Reference Adams, Sanderson and Lindor35). However, in a systematic review of paired biopsy studies only the presence of inflammation of any grade and increased age were independent risk factors for fibrosis progression.
In another study, patients with cirrhosis due to NAFLD were compared to a group of patients with cirrhosis due to the hepatitis C virus(Reference Sanyal, Banas and Sargeant36). Patients from both groups were most likely to die from complications of sepsis. However, patients with hepatitis C virus had greater all-cause and liver-related mortality, while those with NAFLD had greater cardiovascular-related mortality.
Hepatocellular carcinoma is an important cause of death in chronic liver disease, predominantly on a background of cirrhosis(Reference El Serag and Mason37). Both obesity and diabetes are risk factors for the development of hepatocellular carcinoma, and in patients with hepatocellular carcinoma and the metabolic syndrome as the only risk factor, 65% had no significant fibrosis, demonstrating that metabolic syndrome, rather than resultant cirrhosis may predispose to hepatocellular carcinoma development(Reference Paradis, Zalinski and Chelbi38).
What treatments are effective in fatty liver?
Optimal management of non-alcoholic fatty liver disease/non-alcoholic steatohepatitis
NAFLD/NASH is only one aspect of the wider metabolic syndrome and so shares a common pathogenesis and therapeutic approach with other related conditions (e.g. dyslipidaemia, diabetes, hypertension and obesity). Optimum management may be delivered using a multi-disciplinary approach. At our institution, a specialist NASH clinic has been established, employing a multi-disciplinary team comprising hepatologist, diabetologist, dietitian, psychologist and specialist nurse. The clinic aims to achieve sustained health improvements by simultaneously addressing the multi-factorial pathogenesis and risk factors of these patients, attempting to engender lifestyle changes and providing a platform for trialling novel therapies.
Lifestyle measures
Hepatic steatosis, as mentioned above, is closely associated with obesity, IR and diabetes. Strategies to modify these risk factors are therefore likely to improve the condition. Only recently has evidence become available to support this approach. Obese patients with NASH were randomised to lifestyle intervention (behavioural, dietary and exercise changes) or to structured education. Over one year, the intervention group lost an average of 9·3% body weight compared with 0·2% in the control group. Associated with the weight loss in the intervention group was a significant improvement in the histological indices of inflammation(Reference Promrat, Kleiner and Niemeier39). While weight loss is an important target for intervention in the overweight and obese NAFLD/NASH population, it should also be remembered that improving skeletal muscle insulin sensitivity through aerobic exercise is also a worthwhile therapeutic goal. In a recent study, obese sedentary individuals underwent a 4-week programme of aerobic exercise. While hepatic steatosis, as measured by 1H MRS, was shown to have decreased by 21%, body mass had not significantly altered(Reference Johnson, Sachinwalla and Walton40).
Nutritional risk factors and therapies in fatty liver disease
A number of dietary factors have been identified, which may predispose to fatty liver disease either in deficiency or excess. There are relatively few studies assessing the impact of specific diets on NAFLD. Nevertheless, the close pathogenic association between NAFLD and the metabolic syndrome makes diets aimed at modifying cardiovascular risk factors in diabetics and obese patients appropriate for many patients with NAFLD. Certainly, the inclusion of n-3 fatty acids (for example, fish oils and walnuts) MUFA (olive oil), fruit and vegetables, low glycaemic index foods, high-fibre foods in a diet low in saturated fat, simple carbohydrates and sweetened drinks may be recommended. This has been comprehensively reviewed elsewhere(Reference Zivkovic, German and Sanyal41), but the effects of certain dietary lipids, choline-related compounds and sugars are considered further below.
Dietary lipids
Obesity and IR predispose not only to fatty liver as described above but also to CVD (the main cause of morbidity and mortality in NAFLD). The effect of dietary cholesterol intake is unclear with some studies showing high cholesterol intake associated with NAFLD(Reference Musso, Gambino and De Michieli42) and others showing no difference(Reference Cortez-Pinto, Jesus and Barros43). Consumption of diets high in SFA increases the risk of CVD. In a rat model, SFA led to hepatocyte injury associated with endoplasmic reticulum stress(Reference Wang, Wei and Pagliassotti44), while in human subjects, diets high in saturated fat and low in PUFA are associated with NASH, compared to controls with the same body mass(Reference Musso, Gambino and Pacini45).
PUFA include both n-6 and n-3 fatty acids, considered essential for the production of long-chain fatty acids, eicosanoids and PG. The n-3 fatty acid, α-linolenic acid, is required for synthesis of the anti-inflammatory long-chain fatty acids, DHA and EPA, which are also found in fish oils. n-6 fatty acids are required for the synthesis of pro-inflammatory arachidonic acid.
Increasing the ratio of dietary n-6:n-3 fatty acids in rats with sucrose-induced IR improved IR(Reference Ghafoorunissa, Ibrahim and Rajkumar46), while in human subjects, the ratio of dietary n-6:n-3 fatty acids was significantly higher in those with NAFLD compared to healthy controls, when assessed using a dietary history(Reference Cortez-Pinto, Jesus and Barros43).
On the basis of animal data which showed that dietary supplementation with fish oils in rats prevented IR induced by high fat feeding(Reference Storlien, Kraegen and Chisholm47) and reduced hepatic steatosis(Reference Sekiya, Yahagi and Matsuzaka48), interventional studies in human subjects have been conducted. n-3 PUFA supplementation (EPA and DHA at a ratio of 0·9:1·5) in 42 patients with NAFLD led to significant decreases in serum liver enzyme levels, glucose and TAG. Furthermore, there was a reduction in hepatic steatosis as assessed by ultrasound(Reference Capanni, Calella and Biagini49). In patients with NASH, n-3 PUFA (EPA) supplementation over a one-year period led to improvements in lipid indices, serum liver enzyme levels overall and in histological assessment of fibrosis, steatosis and inflammation in the seven patients who had undergone liver biopsy before and after treatment(Reference Tanaka, Sano and Horiuchi50). These findings need confirmation in larger-scale randomised controlled studies, before high-dose fish oil supplementation can be recommended in clinical practice.
The beneficial effects of n-3 PUFA are thought to be exerted by numerous inter-related mechanisms. EPA and DHA are ligands of the nuclear receptor PPARα. Reduced PPARα activity, seen when there are low levels of circulating n-3 fatty acids, may result in a number of processes including: increases in insulin sensitivity and mitochondrial oxidation of fatty acids; reduction of hepatic uptake of NEFA; reduced hepatic lipogenesis (in part by up-regulation of lipogenic transcription factors such as sterol regulatory element binding protein-1); increased lipid export in the form of VLDL, all leading to reduction in hepatic steatosis and reactive oxygen species production(Reference Capanni, Calella and Biagini49, Reference Lee, Gura and Puder51) (Fig. 1).
Choline-related compounds
Choline is an essential nutrient required for numerous physiological processes including the cell membrane manufacture and function, lipid transport and methionine synthesis. Mice fed a choline deficient diet develop hepatic steatosis and steatohepatitis(Reference Walkey, Yu and Agellon52), while those fed a high fat diet were found to have decreased levels of hepatic choline-related compounds(Reference Cobbold, Anstee and Goldin31). Phosphatidylcholine, synthesised from choline, is required for the normal hepatic secretion of VLDL(Reference Yao and Vance53) and decreased phosphatidylcholine leads to the accumulation of intracellular TAG(Reference Noga, Zhao and Vance54). Moreover, 65% of men and women fed a diet deficient in choline developed liver dysfunction (largely steatosis as assessed by MRI)(Reference Fischer, da Costa and Kwock55). A diet rich in choline and betaine was also associated with low levels of serum markers of inflammation, such as C-reactive protein, IL-6 and TNF (reviewed by Zeisel and da Costa(Reference Zeisel and da Costa56)). As such, choline has been considered an essential nutrient by the Institute of Medicine. The choline oxidation product, betaine is also obtained through the diet. It is important as a methyl group donor used in the conversion of homocysteine to methionine and in the synthesis of phosphatidylcholine. Through varied effects on intermediary metabolism, betaine has been shown to reduce hepatic steatosis in animal models, forming the basis of trials to test its efficacy at pharmacological doses, as described later(Reference Abdelmalek, Sanderson and Angulo57).
Sugars
High fructose maize starch is used as a sweetener in many processed foods and carbonated drinks. High intake of soft drinks is associated with an increased risk of NAFLD, partly through increased energy intake and in part through the consumption of advanced glycation end-products which are associated with increased IR and inflammation. Fructose intake is associated with development of NAFLD(Reference Ouyang, Cirillo and Sautin58), with cases of NAFLD consuming over twice as much as the matched control group(Reference Zelber-Sagi, Nitzan-Kaluski and Goldsmith59). Fructose consumption is thought to promote hepatic lipogenesis as glycolysis of fructose is not subject to the rate-limiting step of metabolism by phosphofructokinase, leading to unregulated increase in the acetyl-CoA, a substrate for lipogenesis. Fructose also up-regulates sterol regulatory element-binding protein-1c, increasing de novo lipogenesis(Reference Stanhope and Havel60). Furthermore, fructose consumption increases postprandial hypertriacylglycerolaemia, while reducing glucose concentrations, when subjects consumed a fructose-sweetened drink with every meal for 2 d. Fructose feeding was also observed to reduce postprandial suppression of ghrelin secretion and reducing postprandial leptin and insulin secretion, likely resulting in decreased satiety and increased energy intake(Reference Teff, Elliott and Tschop61).
Pharmacological therapies
Those pharmacological therapies that have been applied to the treatment of NAFLD/NASH in clinical trials have provided encouraging but often mixed results at times. Table 3 highlights recent randomised controlled trials, using liver biopsy features to assess change with treatment. PPARγ binding drugs (thiazolidinediones) show promise for the treatment of NASH, but are associated with overall weight gain(Reference Ratziu, Giral and Jacqueminet62, Reference Aithal, Thomas and Kaye63). Furthermore, there is conflicting evidence of possible increased cardiovascular morbidity and mortality associated with their use. Despite a clear rationale and promising pilot data, betaine failed to demonstrate a consistent beneficial effect(Reference Abdelmalek, Sanderson and Angulo57), while the use of orlistat did not confer any specific benefit, though this study and that of weight loss per se both showed that significant weight loss is associated with histological improvement in NASH(Reference Promrat, Kleiner and Niemeier39, Reference Harrison, Fecht and Brunt64).
Table 3. Selected randomised controlled trials of some pharmacological therapies for non-alcoholic steatohepatitis (NASH)
NAFLD, non-alcoholic fatty liver disease.
An important recent study has compared the efficacy of Pioglitazone to antioxidant (vitamin E) treatment and placebo in histologically proven steatohepatitis (Table 3). While Pioglitazone was found to provide some beneficial effects, a key finding was that vitamin E given at a pharmacological dose (800 IU once daily) rather than thiazolidinediones therapy produced statistically significant improvements in histological steatohepatitis(Reference Sanyal, Chalasani and Kowdley65). Given the lack of targeted therapies that are proven to ameliorate steatohepatitis, it is likely that vitamin E will become part of the standard of care for NAFLD/NASH.
There remains a need to develop further targeted therapies both to address the initiation of steatohepatitis and the progression of liver fibrosis. Based on our developing understanding of disease pathogenesis, a range of agents has been tested in experimental models. One area that has provided promising results is caspase inhibition; however, the effects of long-term use are as yet unknown and to date they have not been studied in clinical trials(Reference Witek, Stone and Karaca66, Reference Anstee, Concas and Kudo67). There is now an emerging literature providing a foundation for evidence-based therapy in NAFLD/NASH. Epidemiological and interventional data, backed by basic science, point to certain dietary constituents (such as fructose and saturated fats) being associated with NASH, and some (n-3 PUFA) being potentially protective against NASH. As CVD and malignancies are the major causes of death in NAFLD/NASH, modification of risk factors for these outcomes (such as smoking, hypertension and obesity) should form a prominent part of management strategies for NAFLD. However, in patients with NAFLD/NASH, liver-related morbidity and mortality also far exceeds that of the general population, so screening for complications of chronic liver disease (hepatocellular carcinoma) and portal hypertension (oesophageal varices) is warranted, particularly where the risk of severe fibrosis or cirrhosis is high.
The importance of fatty liver
The importance of fatty liver has been discussed from the perspective of the epidemiology, pathogenesis, associated morbidity and mortality and treatment options. Hepatic steatosis is present in about a quarter to a third of the adult population in developed countries. Tools are increasingly available to screen for and stage the severity of disease. The pathogenesis is increasingly understood informing the development of treatment strategies. Our understanding of the natural history of the disorder and epidemiological studies of morbidity and mortality have demonstrated that a proportion of patients go on to develop progressive disease, which in absolute terms results in a large number of patients with significant liver disease. This is echoed by a small but increased risk of liver-related mortality and substantial all-cause mortality as a result of the multi-system consequences of the metabolic syndrome. Finally, simple dietary and lifestyle measures, when adhered to, lead to improvements in disease, while an increasing number of specific treatments are being developed. This highlights the importance of the condition and supports the case for increased clinical awareness, targeted service development and basic and clinical research.
Acknowledgements
J. F. L. C. is a Walport Clinical Lecturer at Imperial College London. Q. M. A. and S. T. R. are Imperial College/NIHR Biomedical Research Centre investigators and hold grants from the Medical Research Council and the Wellcome Trust. All authors are grateful to the NIHR Biomedical Facility at Imperial College London for infrastructure support. The authors declare no conflict of interest. J. F. L. C., Q. M. A. and S. D. T. R. wrote the manuscript and approved the final version.
