GI Epidemiology: nonalcoholic fatty liver disease




Background  Nonalcoholic fatty liver disease (NAFLD) is a common diagnosis in clinical practice. Insulin resistance and oxidative stress play an important role in NAFLD development and progression.

Aim  To review the data available on the epidemiology and natural history of NAFLD as well as the risk factors for its development and the areas where future research is necessary.

Results/Conclusions  NAFLD may affect individuals of any age range and race/ethnicity. NAFLD affects one in three adults and one in ten children/adolescents in the United States. Mortality in patients with NAFLD is significantly higher than in the general population of same age and gender with liver-related complications being a common cause of death. Liver-related morbidity and mortality in NAFLD occurs when the disease has progressed to advanced fibrosis and cirrhosis. Further studies are necessary to determine the impact of NAFLD on health-related quality of life and resources utilization, and to the extent to which preventing the development of the metabolic syndrome would prevent NAFLD development and reduce liver-related morbidity and mortality. Lifestyle intervention may improve NAFLD, but medications that increase insulin sensitivity and the antioxidant defenses in the liver deserve evaluation in carefully controlled trials.

Key points

  • Nonalcoholic fatty liver disease (NAFLD) comprises a spectrum of liver pathology including bland steatosis, steatohepatitis, cirrhosis and hepatocellular carcinoma.
  • NAFLD affects a substantial proportion of the general population from several countries.
  • The prevalence and incidence of NAFLD is expected to increase worldwide as the global obesity epidemic grows, and with the trend in developing countries toward a Western lifestyle.
  • Insulin resistance is almost a universal finding in patients with NAFLD, and NAFLD is considered the hepatic manifestation of the metabolic syndrome, which includes central obesity, hyperglycemia, low HDL (high-density lipoprotein) cholesterol, hypertension and hypertriglyceridemia.
  • Improvement of insulin resistance with lifestyle intervention constitutes an essential step in both the treatment and prevention of nonalcoholic fatty liver disease. Medications that increase insulin sensitivity and the antioxidant defenses in the liver hold promise for the treatment of NAFLD.

Clinical summary

Patients may complain of fatigue or malaise and a sensation of fullness or discomfort in the right upper abdomen. Health-related quality of life is significantly diminished due to insulin resistance-associated comorbidities. Hepatomegaly and acanthosis nigricans in children are common physical findings. Patients with “cryptogenic” cirrhosis share many clinical features of patients with NAFLD suggesting that their cryptogenic cirrhosis is in fact the cirrhotic stage of unrecognized NAFLD. Insulin resistance and oxidative stress play a key role in the development and progression of NAFLD. Mild to moderate elevation of serum aminotransferases is the most common and often the only laboratory abnormality found in patients with NAFLD. The aspartate transaminase/alanine transaminase (AST/ALT) ratio is usually less than one, but this ratio increases as fibrosis advances.

Imaging studies, including ultrasonography, computed tomography (CT) scan and magnetic resonance (MR) imaging are sensitive in detecting steatosis, but the grade and stage of disease can be determined only with a liver biopsy. Histological features include steatosis alone or in combination with mixed inflammatory cell infiltration, hepatocyte ballooning and necrosis, Mallory’s hyaline and fibrosis. These histological features are mostly seen in acinar zone 3, although portal-based injury is commonly seen in children.

The diagnosis of NAFLD requires the exclusion of alcohol abuse and other etiologies as the cause of the liver disease. Treatment of patients with NAFLD should focus on the management of associated conditions including obesity, and glucose and lipid abnormalities (Fig. 34.1). Lifestyle intervention with change of diet and increased physical activity is the cornerstone in the management of NAFLD. Medications, including insulin sensitizers and antioxidants, are being evaluated in placebo-controlled trials. Patients with NAFLD with simple steatosis seem to follow a relatively benign course, whereas in others, NAFLD progresses to advanced fibrosis and cirrhosis with its consequent complications of portal hypertension and liver failure. Cirrhotic stage NAFLD constitutes a common indication for liver transplantation to date. As in other types of cirrhosis, cirrhotic-stage NAFLD may be complicated by hepatocellular carcinoma (HCC) [1, 2].

Figure 34.1.

 Diagnosis and treatment algorithm for nonalcoholic fatty liver disease (NAFLD).

Disease definition

NAFLD refers to the accumulation of fat, mainly triglycerides, in hepatocytes so that it exceeds 5% of the liver weight. Primary NAFLD results from insulin resistance and thus frequently occurs as part of the metabolic changes that accompany obesity, type 2 diabetes and dyslipidemia. However, it is important to exclude secondary causes of steatosis (Table 34.1). The histological damage in NAFLD is very similar to that seen in patients with alcoholic liver disease, but NAFLD is by definition not alcohol-induced. Alcohol abuse, hepatotoxic medications and other liver conditions should be ruled out. However, given the high prevalence of obesity, diabetes and dyslipidemia in the general population, NAFLD often coexists with liver diseases of other etiology [3].

Table 34.1.   Causes of nonalcoholic fatty liver disease (NAFLD)
PrimaryObesity, glucose intolerance, type 2 diabetes, hypertriglyceridemia, low HDL (high-density lipoprotein) cholesterol, hypertension
NutritionalProtein-calorie malnutrition, rapid weight loss, gastrointestinal bypass surgery, total parenteral nutrition
DrugsGlucocorticoids, estrogens, tamoxifen, amiodarone, methotrexate, diltiazem, zidovudine, valproate, aspirin, tetracycline, cocaine
MetabolicLipodystrophy, hypopituitarism, dysbetalipoproteinemia, Weber–Christian disease
ToxinsAmanita phalloides mushroom, phosphorus poisoning, petrochemicals, Bacillus cereus toxin
InfectionsHuman immunodeficiency virus, hepatitis C, small bowel diverticulosis with bacterial overgrowth

Incidence and prevalence

The incidence of NAFLD remains unknown because no prospective studies have been conducted. The true prevalence of NAFLD and its different stages remains incompletely defined. The reported prevalence of NAFLD varies based on the information available in a given population and the diagnostic criteria used. Table 34.2 summarizes the results of several studies on the prevalence of NAFLD. Population-based studies provide more accurate figures, but few such studies have been reported to date. Using proton MR spectroscopy, the Dallas Heart Study (a population-based cohort study performed in an ethnically diverse community in the USA) reported that one in three adult Americans have steatosis [3]. The finding indicates that over 70 million adult Americans suffer from NAFLD. In that study [3], 79% of patients with NAFLD had normal aminotransferase levels, and thus, studies using liver enzymes as a surrogate for NAFLD underestimate the prevalence of NAFLD. A high prevalence rate of NAFLD has been reported from other countries. Using liver ultrasonography, a recent population-based cohort study performed in Italy found that one in four or five adults in that country suffer from NAFLD [4]. NAFLD has also reached epidemic proportions among populations typically considered at “low risk” for this liver condition, with a prevalence in China and Japan of 15% and 14%, respectively, among adults. The clinical implications of this alarming prevalence of NAFLD are derived from the fact that this liver condition may progress to end-stage liver disease and liver cancer [5, 6].

Table 34.2.   Selected studies on prevalence of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH)
Author (year)StudyDiagnostic methodCountryNo. of individuals screened Prevalence of NAFLD (%) Prevalence of NASH (%)
  1. aPediatric series. bObese children. ND, not determined.

Browning (2004)Population-basedMR spectroscopyUSA228731ND
Bedogni (2005)Population-basedUltrasonographyItaly59823ND
Fan (2005)Population-basedUltrasonographyChina317515ND
Nomura (1988)Population-basedUltrasonographyJapan257414ND
Clark (2003)Population-basedAminotransferasesUSA15 6765.4ND
Ruhl (2003)Population-basedAminotransferasesUSA57242.8ND
Jimba (2005)Health evaluationUltrasonographyJapan195029ND
Hamaguchi (2005)Health evaluationUltrasonographyJapan440118ND
Park (2006)Health evaluationUltrasonographySouth Korea664816ND
Hultcrantz (1986)Hospital seriesLiver biopsySweden14939ND
Lee (1989)Hospital seriesLiver biopsyUSA543ND9
Nonomura (1992)Hospital seriesLiver biopsyJapan561ND1
Byron (1996)Hospital seriesLiver biopsyUSA1226ND11
Daniel (1999)Hospital seriesLiver biopsyUSA815132
Berasain (2000)Hospital seriesLiver biopsySpain1075ND16
Hilden (1977)Autopsy seriesLiver biopsySweden50324ND
Ground (1982)Autopsy seriesLiver biopsyUSA42316ND
Wanless (1990)Autopsy seriesLiver biopsyCanada207296
El-Hassan (1992)OutpatientsUltrasonography, CTSaudi Arabia142510ND
Lonardo (1997)OutpatientsUltrasonographyItaly36320ND
Araujo (1998)OutpatientsUltrasonographyBrazil21733.5ND
Omagari (2002)OutpatientsUltrasonographyJapan34329ND
Luyckx (1998)Bariatric surgeryLiver biopsyBelgium52874ND
Silverman (1990)Bariatric surgeryLiver biopsyUSA1008636
Dixon (2001)Bariatric surgeryLiver biopsyAustralia1057125
Beymer (2003)Bariatric surgeryLiver biopsyUSA488533
Spaulding (2003)Bariatric surgeryLiver biopsyUSA488856
Mathurin (2006)Bariatric surgeryLiver biopsyFrance167ND14.4
Franzese (1997)a,bOutpatientsUltrasonographyItaly7253ND
Tominaga (1995)aHealth evaluationUltrasonographyJapan8103ND
Schwimmer (2006)aAutopsy seriesLiver biopsyUSA7429.6 (38 among obese)3

Population-based studies provide better estimates of the prevalence of NAFLD in the general population compared with autopsy studies, hospital series or studies performed exclusively in obese populations (Table 34.2). The prevalence of NAFLD among children is unknown, but some data indicate that 2.6–9.6% of children have NAFLD, increasing up to 38–53% among obese children (Table 34.2).

Risk factors for disease

NAFLD may affect any age and ethnic group. The prevalence of NAFLD among adults in the USA seems to be different among different ethnic groups, affecting 45% of Hispanic people, 33% of White people and 24% of Black people. The prevalence is significantly higher in White men (42%) than in White women (24%). There is no gender difference in prevalence among Hispanic people or Black people [7]. In children and teens, the prevalence of NAFLD again seems to be different among the different ethnic groups, with the highest prevalence among Hispanic people and the lowest among Black children [8]. Differences in body fat distribution and body composition among the different ethnic groups may partially explain the racial differences in prevalence. For instance, Hispanic people have a higher proportion of body fat and higher waist to hip ratio than their taller counterparts [9]. Similarly, Asian people have a higher proportion of visceral fat and a lower proportion of lean body mass than White subjects with the same body mass index (BMI) [10].

The central (or upper body) obesity phenotype is associated with increased intra-abdominal (or visceral) fat. Visceral adipose tissue has greater lipolytic potential than subcutaneous adipose tissue, and the release of free fatty acids (FFA) from visceral fat directly into the portal circulation creates a “first-pass” effect [11]. Increased FFA concentrations, in turn, are considered a major mediator of insulin resistance. In contrast, FFA flux and concentrations in individuals with predominantly lower body obesity tend to be normal, regardless of BMI [11]. Therefore, patients with central obesity are characteristically insulin resistant, and more commonly present with NAFLD compared with patients having lower-body obesity [12].

Besides central obesity, type 2 diabetes, dyslipidemia and hypertension are risk factors for the development of NAFLD. NAFLD, however, can also precede the development of these other comorbidities [5].

Environmental factors and lifestyle-related factors such as reduced physical activity and high-fat diets are well-known influences for the development of insulin resistance-associated comorbidities and NAFLD. The genetic predisposition for the development of central obesity and type 2 diabetes undoubtedly plays a role in the development of NAFLD, although family studies and studies specifically addressing the genetic susceptibility for NAFLD development are lacking.

Natural history and mortality

Changes in fibrosis stage have been specifically evaluated in four independent series (Table 34.3). Overall, fibrosis progresses over time, but it remains stable for a number of years in many cases and it may actually improve spontaneously in some cases [6, 13–15]. Higher BMI and greater insulin resistance or the presence of type 2 diabetes are risk factors for a higher rate of fibrosis progression [6, 15]. As fibrosis develops and progresses over time, other features of NAFLD, including steatosis, inflammation and ballooning of hepatocytes, significantly improve or disappear [15]; thus, liver biopsy features other than fibrosis severity may not be useful to predict the long-term prognosis in an individual patient with NAFLD. Furthermore, the histological features of NAFLD that create the basis for the histological diagnosis of nonalcoholic steatohepatitis (NASH) (i.e., inflammation and hepatocyte ballooning) are unequally distributed throughout the liver parenchyma, with liver biopsy resulting in misdiagnosis in some patients [16]. Also, aminotransferases, when elevated, improve or normalize spontaneously over time despite fibrosis progression [15].

Table 34.3.   Changes in fibrosis stage evaluated in studies with sequential liver biopsy in nonalcoholic fatty liver disease (NAFLD)
Author (year)No. of patientsAverage time interval (years) between biopsies (range)Progressed n (%)Stable n (%)Improved n (%)
Harrison (2003)225.7 (1.4–15.7)7 (32)11 (50)4 (18)
Fassio (2004)224.3 (3–14.3)7 (32)11 (50)4 (18)
Adams (2005)1033.2 (0.7–21.3)38 (37)35 (34)30 (29)
Ekstedt (2006)7013.8 (10.3–16.3)29 (41)30 (43)11 (16)

Studies evaluating the long-term prognosis of patients with NAFLD are summarized in Table 34.4. Overall, the disease progresses slowly over many years or decades, but the prognosis is different across the different stages of NAFLD. Patients with simple, bland steatosis appear to have a more benign prognosis. For instance, a Danish study of a cohort of 109 predominantly morbidly obese subjects followed for nearly 17 years found the incidence of cirrhosis to be <1% [17]. During follow-up, a quarter of patients died but the survival curve of the general population fell within the 95% confidence interval of the survival curve of patients with bland steatosis. In that study [17], the patient who developed cirrhosis was the only one who died from liver-related causes. Conversely, patients with cirrhotic stage NASH have a worse prognosis, as demonstrated in three recent studies [18–20]. In those studies, 9–26% of patients died within 4–10 years of follow-up, with most causes of death related to end-stage liver disease.

Table 34.4.   Studies on long-term prognosis of nonalcoholic fatty liver disease (NAFLD)
Author (year) DiagnosisanCirrhosis prevalence (%)bNo. of liver-related deaths (%)No. of deaths overall (%)Average follow-up (years)
  1. aNAFLD denotes the inclusion of both patients with simple steatosis and patients with nonalcoholic steatohepatitis (NASH).

  2. bCirrhosis prevalence includes all patients diagnosed with cirrhosis at both baseline and during follow-up.

Teli (1995)Bland steatosis400014 (35)9.6
Dam-Larsen (2004)Bland steatosis10911 (0.9)27 (24.8)16.7
Matteoni (1999)NAFLD98209 (9)48 (49)8.3
Adams (2005)NAFLD42057 (1.7)53 (12.6)7.6
Ekstedt (2006)NAFLD1297.82 (1.6)26 (20.2)13.7
Lee (1989)NASH3916.31 (3)10 (26)3.8
Powell (1990)NASH4271 (2)2 (5)4.5
Evans (2002)NASH26404 (15)8.7
Hui (2004)Cirrhotic-stage NASH231005 (21)6 (26)5.0
Hashimoto (2005)NASH with septal fibrosis or cirrhosis89486 (6.7)8 (9)3.7
Sanyal (2006)Cirrhotic-stage NASH15210022 (14.5)29 (19.1)10

Overall, a diagnosis of NAFLD is associated with a shorter survival than expected for the general population of the same age and gender, as recently demonstrated in two independent studies [5, 6]. A community-based study performed in the USA included 420 patients with NAFLD and found liver-related complications to be the third most common cause of death among NAFLD patients compared with the 13th most common cause of death in the general population [5]. This indicates that complications of end-stage liver disease contribute importantly to mortality in patients with NAFLD. Patients dying from liver-related causes were those with more advanced NAFLD [5], confirming observations of smaller studies [17–20]. Impaired fasting glucose or diabetes, older age and presence of cirrhosis are risk factors independently associated with a higher mortality in NAFLD [5].

Interestingly, a recent Swedish study of 129 patients presenting with abnormal liver enzymes found a significantly higher mortality among patients with NAFLD compared with the general population of the same age and gender after almost 14 years of follow-up [6]. Again, liver-related complications were the third most common cause of death among NAFLD patients, with cardiovascular disease and extrahepatic malignancy being the first and second most common causes of death, respectively.

The potential for NAFLD to result in end-stage liver disease is further highlighted by some data suggesting that NAFLD underlies a substantial proportion of cases of cryptogenic cirrhosis [21]. Of patients with cryptogenic cirrhosis, 50–73% have a BMI in the obese category or suffer from diabetes. The prevalence of NAFLD as an unrecognized cause of cryptogenic cirrhosis is probably underestimated because some nondiabetic, nonobese (i.e., BMI <30) patients may suffer from central obesity and/or dyslipidemia, which may be the only risk factor(s) for NAFLD and have not been consistently measured in series of cryptogenic cirrhosis. Further, the presence of NAFLD increases disease severity and progression in other liver diseases including chronic hepatitis C infection, alcoholic liver disease and hemochromatosis [3].

Quality of life

The impact of NAFLD on health-related quality of life is currently been evaluated. Several studies have found a significant detrimental impact on health-related quality of life of the several comorbidities that conform the metabolic syndrome and that often cluster with NAFLD.


There are no studies of measures aimed at preventing NAFLD development. However, preventing the development of insulin resistance and its clinical manifestations (i.e., the metabolic syndrome) is expected to prevent NAFLD development. Weight gain and obesity resulting from a more sedentary lifestyle and high-fat diets seem to be key factors in the development of insulin resistance and NAFLD [1]. Thus, achieving and maintaining appropriate weight control would be expected to prevent the development of NAFLD, as would the treatment of glucose and lipid abnormalities. This is further supported by data from the diabetes prevention program in the USA [22], demonstrating that both lifestyle intervention and the insulin-sensitizing drug, metformin, significantly reduce the development of the metabolic syndrome, which, intuitively, would prevent the development of NAFLD.

Issues in epidemiology knowledge

There is a relative scarcity of NAFLD prevalence data available from population-based studies. There are no data on the change in prevalence of NAFLD within a population over time, and there are no data on incidence of NAFLD. The lack of a diagnostic test or combination of tests with 100% accuracy precludes firm conclusions about the incidence and prevalence of NAFLD, and its different stages, in the general population. Liver enzymes are insensitive and nonspecific for chronic liver disease. Imaging techniques such as ultrasonography and CT scan may provide false negatives. More sensitive techniques, including MR imaging and spectroscopy, are hindered by expense and lack of feasibility in large populations. Liver biopsy has been considered as the gold standard, but is limited by sampling and interpretation error besides its cost and impractical applicability in population-based studies. Furthermore, unless uniform data become available, estimates of the prevalence and incidence of NAFLD over a given time period will probably be affected by increased awareness of the disease.

Recommendations for future studies

Further population-based studies are necessary to determine the true prevalence and the impact on health-related quality of life of NAFLD. Prospective studies with long-term follow-up will better define the natural history of NAFLD and its incidence in specific populations. Genetic studies are necessary to determine to what extent the genetic background predisposes to NAFLD development and progression to advanced liver disease. Carefully controlled clinical trials will better define the impact of lifestyle intervention and pharmacotherapy on NAFLD [23].


With the increasing prevalence of obesity, type 2 diabetes and the metabolic syndrome in the general population, NAFLD has become a common diagnosis in clinical practice of several medical specialties. Bland steatosis remains stable for a number of years and will probably never progress in many cases, with most liver-related morbidity and mortality observed in those patients whose disease progresses to advanced fibrosis and cirrhosis. Further studies are necessary to determine the impact of NAFLD on health-related quality of life and resources utilization as well as the extent to which preventing the development of the metabolic syndrome would prevent NAFLD development and reduce liver-related morbidity and mortality.