Abstract

These recommendations are based on the following: (1) a formal review and analysis of the recently published world literature on the topic [Medline search up to June 2011]; (2) the American College of Physicians' Manual for Assessing Health Practices and Designing Practice Guidelines;1 (3) guideline policies of the three societies approving this document; and (4) the experience of the authors and independent reviewers with regards to NAFLD. Intended for use by physicians and allied health professionals, these recommendations suggest preferred approaches to the diagnostic, therapeutic and preventive aspects of care. They are intended to be flexible and adjustable for individual patients. Specific recommendations are evidence-based wherever possible, and when such evidence is not available or inconsistent, recommendations are made based on the consensus opinion of the authors. To best characterize the evidence cited in support of the recommendations, the AASLD Practice Guidelines Committee has adopted the classification used by the Grading of Recommendation Assessment, Development, and Evaluation (GRADE) workgroup with minor modifications (Table 1).2 The strength of recommendations in the GRADE system is classified as strong (1) or weak (2). The quality of evidence supporting strong or weak recommendations is designated by one of three levels: high (A), moderate (B) or low-quality (C).2 This is a practice guideline for clinicians rather than a review article and interested readers can refer to several comprehensive reviews published recently.3-8 NAFLD, Nonalcoholic Fatty Liver Disease; NAFL,Nonalcoholic Fatty Liver; NASH, Nonalcoholic Steatohepatitis; T2DM, Type 2 Diabetes Mellitus; AST, Aspartate Aminotransferase; ALT, Alanine Aminotransferase; HOMA,Homeostatic Model Assessment; RCT; Randomized Controlled Trial; PIVENS: Pioglitazone versus Vitamin E versus Placebo for the Treatment of Non-diabetic patients with Nonalcoholic steatohepatitis; TONIC; Treatment of Nonalcoholic Fatty Liver Disease in Children; NAS, NAFLD Activity Score; CK18; Cytokeratin 18 Fragments; ELF, Enhanced Liver Fibrosis Panel; TZD; Thiazolidinediones; UDCA: Ursodeoxycholic Acid; ANA; Anti Nuclear Antibody; ASMA: Anti Smooth Muscle Antibody; US; Ultrasound; CT: Computerized Tomography; MR; Magnetic Resonance. The definition of nonalcoholic fatty liver disease (NAFLD) requires that (a) there is evidence of hepatic steatosis, either by imaging or by histology and (b) there are no causes for secondary hepatic fat accumulation such as significant alcohol consumption, use of steatogenic medication or hereditary disorders (Table 2). In the majority of patients, NAFLD is associated with metabolic risk factors such as obesity, diabetes mellitus, and dyslipidemia. NAFLD is histologically further categorized into nonalcoholic fatty liver (NAFL) and nonalcoholic steatohepatitis (NASH) (Table 3). NAFL is defined as the presence of hepatic steatosis with no evidence of hepatocellular injury in the form of ballooning of the hepatocytes. NASH is defined as the presence of hepatic steatosis and inflammation with hepatocyte injury (ballooning) with or without fibrosis. The incidence of NAFLD has been investigated in a limited number of studies. Two Japanese studies9, 10 reported an incidence rate of 31 and 86 cases of suspected NAFLD per 1,000 person-years respectively, whereas another study from England showed a much lower incidence rate of 29 cases per 100,000 person-years.11 More studies are needed to better understand the incidence of NAFLD across different age, ethnic, and geographic groups. The reported prevalence of NAFLD varies widely depending on the population studied and the definition used. The prevalence of histologically-defined NAFLD was 20% and 51% in two different studies comprised of potential living liver donors.12, 13 The reported prevalence of NAFLD when defined by liver ultrasound ranged between 17% and 46% depending on the population studied.4 In a study consisting of nearly 400 middle aged individuals, the prevalence of NAFLD defined by ultrasonography was 46% and the prevalence of histologically confirmed NASH was 12.2%.14 In the Dallas Heart Study, when assessed by MR spectroscopy the prevalence of NAFLD in the general population was 31%.15 The prevalence of suspected NAFLD when estimated using aminotransferases alone without imaging or histology ranged between 7% and 11%, but aminotransferases can be normal in individuals with NAFLD.4 In summary, estimates of the worldwide prevalence of NAFLD ranges from 6.3% to 33% with a median of 20% in the general population, based on a variety of assessment methods.4 On the other hand, the estimated prevalence of NASH is lower, ranging from 3 to 5%.4 The prevalence of NASH cirrhosis in the general population is not known. Obesity is a common and well documented risk factor for NAFLD. Both excessive BMI and visceral obesity are recognized risk factors for NAFLD. In patients with severe obesity undergoing bariatric surgery, the prevalence of NAFLD can exceed 90% and up to 5% of patients may have unsuspected cirrhosis.4, 16-20 There is a very high prevalence of NAFLD in individuals with type 2 diabetes mellitus (T2DM).4 An ultrasonographic study of patients with T2DM showed a 69% prevalence of NAFLD.21 In another study, 127 of 204 diabetic patients displayed fatty infiltration on ultrasound, and 87% of the patients with fatty infiltration who consented to biopsy had histologic confirmation of NAFLD.22 High serum triglyceride levels and low serum HDL levels are very common in patients with NAFLD. The prevalence of NAFLD in individuals with dyslipidemia attending lipid clinics was estimated to be 50%.23 Age, gender and ethnicity are also associated with a differential prevalence for NAFLD.4 A number of studies have shown that the prevalence of NAFLD increases with age.24-28 The likelihood of disease progression to advanced fibrosis or mortality increases in older patients with NAFLD.29-31 Many recent studies have reported that male gender is a risk factor for fatty liver disease.4 For example, in a study of 26,527 subjects undergoing medical checkups, the prevalence of NAFLD was 31% in men and 16% in women.32 Compared to non-Hispanic whites, Hispanic individuals have significantly higher and non-Hispanic blacks have significantly lower prevalence of NAFLD.15, 33-35 The prevalence of NAFLD in American-Indian and Alaskan-Native populations appears lower, ranging from 0.6% to 2.2%, although the lack of histologic definition makes it likely that is an underestimate.36, 37 There are data to suggest that hypothyroidism, hypopituitarism, hypogonadism, sleep apnea, and polycystic ovary syndrome independent of obesity are important risk factors for the presence of NAFLD (Table 4).3 The evolution of hepatic histologic changes in patients with NAFL and NASH has been investigated by several studies, but these generally included smaller number of patients and had relatively modest duration of follow-up.4, 7 Nonetheless, it is generally agreed that patients with simple steatosis have very slow, if any, histological progression, while patients with NASH can exhibit histological progression to cirrhotic-stage disease.4, 7 The long term outcomes of patients with NAFLD and NASH have been reported in several studies.31, 38-45 Their detailed discussion is beyond the scope of this guideline, but their findings can be summarized as follows; (a) patients with NAFLD have increased overall mortality compared to matched control populations, (b) the most common cause of death in patients with NAFLD, NAFL and NASH is cardiovascular disease, and (c) patients with NASH (but not NAFL) have an increased liver-related mortality rate. Another piece of indirect evidence that supports the progressive nature of NASH is in the features of cryptogenic cirrhosis which is closely related to NAFLD.46, 47 Patients with cryptogenic cirrhosis have disproportionately high prevalence of metabolic risk factors (T2DM, obesity, metabolic syndrome) typical of patients with NAFLD, their liver biopsies frequently show one or more features of NASH, and studies have demonstrated the loss of histological features of NASH with the development of cirrhosis.4, 7, 46, 47 Patients with NAFLD are at increased risk for HCC, but this risk is likely limited to those with advanced fibrosis and cirrhosis.48-53 Several studies investigated the natural history of NASH cirrhosis in comparison to patients with hepatitis C cirrhosis.54-57 One large prospective US-based study55 observed a lower rate of decompensation and mortality in patients with NASH cirrhosis as compared to patients with hepatitis C cirrhosis. However, a more recent international study56 of 247 NAFLD patients with advanced fibrosis and cirrhosis followed over a mean duration of 85.6 ± 54.5 months showed an overall 10-year survival of 81.5% that was not different from matched patients with hepatitis C cirrhosis. Importantly, both studies have shown that patients with NASH cirrhosis are at significantly lower risk for HCC than patients with hepatitis C cirrhosis.55, 56 By definition, NAFLD indicates the lack of any evidence of ongoing or recent consumption of significant quantities of alcohol. However, the precise definition of significant alcohol consumption in patients with suspected NAFLD is uncertain. A recent consensus meeting58 concluded that, for NASH clinical trials candidate eligibility purposes, significant alcohol consumption be defined as >21 drinks per week in men and >14 drinks per week in women over a 2-year period prior to baseline liver histology. Furthermore, this group recommended that validated questionnaires should be used to quantify the amount of alcohol consumption in the context of clinical trials. The definition of significant alcohol consumption in the published NAFLD literature has been inconsistent and ranged from > 1 alcoholic drink (∼ 10 grams of alcohol per one drink unit) per day to > 40 grams per day, and published studies have not always used gender-specific definitions.59 If self-reported alcohol consumption details are not consistent with clinical suspicion when evaluating a patient with suspected NAFLD, confirmation with a family member or a close friend should be considered. Recommendation 1. Ongoing or recent alcohol consumption > 21 drinks on average per week in men and > 14 drinks on average per week in women is a reasonable definition for significant alcohol consumption when evaluating patients with suspected NAFLD in clinical practice. (Strength – 2, Quality - C) Some patients undergoing thoracic and abdominal imaging for reasons other than liver symptoms, signs or biochemistry may demonstrate unsuspected hepatic steatosis. While this phenomenon is not uncommon in clinical practice, studies have not systematically examined the characteristics or natural history of NAFLD in this patient population. Recommendations 2. When patients with unsuspected hepatic steatosis detected on imaging have symptoms or signs attributable to liver disease or have abnormal liver biochemistries, they should be evaluated as though they have suspected NAFLD and worked-up accordingly. (Strength – 1, Evidence -A) 3. In patients with unsuspected hepatic steatosis detected on imaging who lack any liver-related symptoms or signs and have normal liver biochemistries, it is reasonable to assess for metabolic risk factors (e.g., obesity, glucose intolerance, dyslipidemia) and alternate causes for hepatic steatosis such as significant alcohol consumption or medications. (Strength – 1, Evidence -A) 4. In patients with unsuspected hepatic steatosis detected on imaging who are asymptomatic and have normal liver biochemistries, a liver biopsy cannot be recommended. (Strength – 1, Evidence -B) It can be argued that there should be systematic screening for NAFLD, at least among higher-risk individuals attending diabetes and obesity clinics. However, at present there are significant gaps in our knowledge regarding the diagnosis, natural history, and treatment of NAFLD. As liver biochemistries can be within normal ranges in patients with NAFLD and NASH, they may not be sufficiently sensitive to serve as screening tests, whereas liver ultrasound is potentially more sensitive but it is expensive and cumbersome as a screening test. Recommendation 5. Screening for NAFLD in adults attending primary care clinics or high-risk groups attending diabetes or obesity clinics is not advised at this time due to uncertainties surrounding diagnostic tests and treatment options, along with lack of knowledge related to the long-term benefits and cost-effectiveness of screening. (Strength – 1, Evidence -B) Anecdotal experience and some published studies suggest familial clustering and heritability of NAFLD,60-63 but conclusive studies are lacking. In a retrospective cohort study, Willner et al. observed that 18% of patients with NASH have a similarly affected first degree relative.61 A small familial aggregation study observed that patients with NAFLD have a significantly higher number of first degree relatives with cirrhosis and a trend towards familial clustering of NAFLD or cryptogenic cirrhosis than matched healthy controls.62 In another familial aggregation study63 of overweight children with and without NAFLD, after adjusting for age, gender, race, and BMI, the heritability of MR-measured liver fat fraction was 0.386, and fatty liver was present in 18% of family members of children with NAFLD despite normal ALT and lack of obesity. Recommendation 6. Systematic screening of family members for NAFLD is currently not recommended. (Strength – 1, Evidence - B) The diagnosis of NAFLD requires that (a) there is hepatic steatosis by imaging or histology, (b) there is no significant alcohol consumption, (c) there are no competing etiologies for hepatic steatosis, and (d) there are no co-existing causes for chronic liver disease. Common alternative causes of hepatic steatosis are significant alcohol consumption, hepatitis C, medications, parenteral nutrition, Wilson's disease, and severe malnutrition (Table 2). When evaluating a patient with newly suspected NAFLD, it is important to exclude co-existing etiologies for chronic liver disease including hemochromatosis, autoimmune liver disease, chronic viral hepatitis, and Wilson's disease.3 Mildly elevated serum ferritin is common in patients with NAFLD and it does not necessarily indicate increased iron stores.3, 64 Elevated serum ferritin and transferrin saturation in patients with suspected NAFLD should lead to testing for genetic hemochromatosis. Mutations in the HFE gene occur with variable frequency in patients with NAFLD and their clinical significance is unclear.64 One should consider a liver biopsy to assess hepatic iron concentration and to exclude significant hepatic injury and fibrosis in a patient with suspected NAFLD with elevated serum ferritin and a homozygote or compound heterozygote C282Y mutation in the HFE gene.65 Elevated serum autoantibodies are common in patients with NAFLD and are generally considered to be an epiphenomenon.3 In a recently published large study from the NASH Clinical Research Network, positive serum autoantibodies, defined as ANA > 1:160 or ASMA >1:40 were present in 21% of patients with well-phenotyped NAFLD and were not associated with more advanced histologic features.66 Recommendations 7. When evaluating a patient with suspected NAFLD, it is essential to exclude competing etiologies for steatosis and co-existing common chronic liver disease. (Strength – 1, Evidence - A) 8. Persistently high serum ferritin and increased iron saturation, especially in the context of homozygote or heterozygote C282Y HFE mutations may warrant a liver biopsy. (Strength – 1, Evidence - B) 9. High serum titers of autoantibodies in association with other features suggestive of autoimmune liver disease (very high aminotransferases, high globulin) should prompt a more complete work-up for autoimmune liver disease. (Strength – 1, Evidence - B) The natural history of NAFLD is fairly dichotomous – NAFL is generally benign whereas NASH can progress to cirrhosis, liver failure, and liver cancer. Existing dogma posits that liver biopsy is the most reliable approach for identifying the presence of steatohepatitis and fibrosis in patients with NAFLD, but it is generally acknowledged that biopsy is limited by cost, sampling error, and procedure-related morbidity and mortality. Serum aminotransferase levels and imaging tests such as ultrasound, CT, and MR do not reliably assess steatohepatitis and fibrosis in patients with NAFLD. Therefore, there has been significant interest in developing clinical prediction rules and non-invasive biomarkers for identifying steatohepatitis in patients with NAFLD,7 but their detailed discussion is beyond the scope of this practice guideline. The presence of metabolic syndrome is a strong predictor for the presence of steatohepatitis in patients with NAFLD3, 7, 67-69 and may be used to best identify patients with persistently abnormal liver biochemistries who would benefit diagnostically and prognostically from a liver biopsy. There has been intense interest in non-invasive methods to identify advanced fibrosis in patients with NAFLD7;these include the NAFLD Fibrosis Score70, Enhanced Liver Fibrosis (ELF) panel70 and transient elastography. The NAFLD Fibrosis Score is based on six readily available variables (age, BMI, hyperglycemia, platelet count, albumin, AST/ALT ratio) and it is calculated using the published formula (http://nafldscore.com). In a meta-analysis of 13 studies consisting of 3,064 patients,7 NAFLD Fibrosis Score has an AUROC of 0.85 for predicting advanced fibrosis (i.e., bridging fibrosis or cirrhosis) and a score 0.676 had 67% sensitivity and 97% specificity to identify the presence of advanced fibrosis. The ELF panel consists of plasma levels of three matrix turnover proteins (hyaluronic acid, TIMP-1, and PIIINP) had an AUROC of 0.90 with 80% sensitivity and 90% specificity for detecting advanced fibrosis.71 Circulating levels of cytokeratin-18 (CK18) fragments have been investigated extensively as novel biomarkers for the presence of steatohepatitis in patients with NAFLD.7, 72 Wieckowska et al., measured CK18 fragments in plasma that had been obtained from 44 consecutive patients with suspected NAFLD at the time of liver biopsy, and correlated the findings with hepatic immunohistochemistry data.70 Plasma CK18 fragments were markedly increased in patients with NASH compared with patients with simple steatosis or normal biopsies (median 765.7 U/L versus 202.4 U/L or 215.5 U/L, respectively; P 5% body weight improved steatosis, whereas individuals with ≥ 9% weight loss had significant improvement in steatosis, lobular inflammation, ballooning, and NAS. A number of recent studies used MR spectroscopy to assess changes in hepatic fat in response to lifestyle modification. The results from these studies using a variety of interventions, either by diet alone81, 83, 84, 89, 92, 93 or in combination with different exercise prescriptions,82, 85-88, 92, 94 have consistently reported a significant reduction in liver fat by an average of ∼40% (ranging from 20% to 81%). The degree of hepatic fat reduction was proportional to the intensity of the lifestyle intervention and generally required a body weight loss between ∼5 to 10%.82, 88, 92 The effect of exercise without dietary modification on hepatic steatosis was investigated in four studies using MR spectroscopy.102-105 Exercise programs consisted of 2-3 sessions a week of 30-60 minutes over a period of 6 to 12 weeks. In all but one study101 liver fat content diminished without a significant change in body weight. Recommendations 16. Weight loss generally reduces hepatic steatosis, achieved either by hypocaloric diet alone or in conjunction with increased physical activity. (Strength – 1, Evidence - A) 17. Loss of at least 3-5% of body weight appears necessary to improve steatosis, but a greater weight loss (up to 10%) may be needed to improve necroinflammation. (Strength – 1, Evidence - B) 18. Exercise alone in adults with NAFLD may reduce hepatic steatosis but its ability to improve other aspects of liver histology remains unknown. (Strength – 1, Evidence - B) Several studies investigated the effect of metformin on aminotransferases and liver histology in patients with NASH. Early small, open-label studies demonstrated a reduction in insulin resistance and aminotransferases106-108 but no significant improvement in liver histology.107, 108 An open-label trial consisting of 110 patients with NASH received either metformin 2 grams/day (55 patients), vitamin E 800 IU/day (28 patients) or dietary-induced weight loss (27 patients) for 12 months.109 Aminotransferases improved more with metformin than with vitamin E or diet alone. However, there was only a modest improvement in hepatic steatosis and inflammation in the subset of 17 patients undergoing paired liver biopsies with metformin treatment. In a 48-week open-label study in 26 patients, metformin improved NASH activity in only 30% of patients, although interpretation of the study was confounded by a significant weight loss in the responders (19% lost more than 10 kilograms).110 Haukeland et al.112 reported a similar lack of efficacy in a larger (n=48) randomized control trial (RCT) of metformin vs. placebo with a similar dietary and exercise intervention in both groups. Other studies also failed to show major benefit for metformin on hepatic insulin sensitivity, aminotransferases111-116 or liver histology.111, 113, 116 A recent meta-analysis4 concluded that 6-12 months of metformin plus lifestyle intervention did not improve aminotransferases or liver histology, compared with lifestyle intervention alone, independently of metformin dose or the presence of diabetes. Recommendation 19. Metformin has no significant effect on liver histology and is not recommended as a specific treatment for liver disease in adults with NASH. (Strength – 1, Evidence - A) Several studies investigated the effect of pioglitazone and rosiglitazone on aminotransferases and liver histology in adults with NASH. In an early uncontrolled open-label study117 in 22 subjects with biopsy-proven NASH, rosiglitazone improved aminotransferases and hepatic steatosis, ballooning and inflammation scores, but not fibrosis. But in a subsequent RCT, Ratziu et al.118 observed that rosiglitazone improved aminotransferases and hepatic steatosis, but not necroinflammation or fibrosis and its two-year open-label extension phase also showed similar results.119 Belfort et al.120 conducted a RCT of pioglitazone (45 mg/day) in patients with NASH who had impaired glucose tolerance or T2DM. Although there was a significant weight gain (2.5 ± 0.5 kg) with pioglitazone, it significantly improved aminotransferases, steatosis, ballooning, and inflammation. The NAS improved with pioglitazone in 73% compared to 24% of placebo-treated patients (P<0.001) and there was a trend towards improvement in fibrosis among patients randomized to pioglitazone (P=0.08). Aithal et al.121 performed a RCT of lifestyle intervention with either pioglitazone 30 mg/day or placebo for 12 months in a total of 74 patients with NASH. While steatosis did not improve significantly compared to placebo, hepatocellular injury and fibrosis improved significantly.1210 The PIVENS122 study is a large multicenter RCT that randomized 247 non-diabetic patients with NASH to pioglitazone (30 mg/day), vitamin E (800 IU/day), or placebo for 24 months. The primary endpoint was an improvement in NAS ≥ 2 points with at least 1 point improvement in hepatocellular ballooning and 1-point improvement in either the lobular inflammation or steatosis score, and no increase in the fibrosis score.122 It was achieved in 19% in the placebo group compared to 34% in the pioglitazone group (P=0.04 vs. placebo) and 43% in the vitamin E group (P=0.001 vs. placebo).122 Because this study consisted of two primary comparisons (pioglitazone vs. placebo and vitamin E vs. placebo), a P-value of 0.025 was considered to be significant a priori. Therefore, although there were histological benefits associated with pioglitazone, this study concluded that pioglitazone did not meet the primary end point. However, resolution of NASH, a key secondary end point, was achieved in significantly higher number of pati