Cause of NAFLD: obesity and abnormalities in adipose tissue
Although both NAFLD and the MetS can occur in nonobese subjects, the prevalence of NAFLD is markedly increased in obesity as is that of the MetS. In the third National Health and Nutrition Examination Survey (NHANES), the prevalence of NAFLD averaged 7.5% and 6.7% in normal-weight men and women but was 57 and 44% in persons with a body mass index (BMI) >35 kgm−2 . Asians have a higher prevalence of NAFLD than Caucasians for a given BMI.For example, the mean BMI in 6905 Chinese adults with ultrasound-diagnosed NAFLD was 23.6 kgm−2 as compared to 21.5 kgm−2 in those without NAFLD . Obesity is a true cause of NAFLD and hepatic insulin resistance as weight loss rapidly reverses both (see  for review). Bariatric surgery induced remission regresses histologic changes, at least steatosis and ballooning, in NASH .
In obese subjects, who have NAFLD, adipose tissue is characterized by multiple alterations such as hypoxia (see  for review), increased infiltration of macrophages surrounding dead adipocytes and expression of chemokines and ro-inflammatory cytokines , fibrosis as well as deficiency of adiponectin . Inflamed adipose tissue is insulin resistant leading to increased FFA flux to the liver especially under conditions mimicking the postprandial state . In obese subjects, a decrease in serum adiponectin is likely to contribute to fat accumulation in the liver and may also induce hepatic inflammation and insulin resistance (see  for review). Changes in gut microbiota have been suggested to contribute to the pathogenesis of both obesity and NAFLD syndrome (see  for review). The change in intestinal bacterial flora is thought to alter the permeability of gut mucosa to inducers of inflammation such as endotoxin. In addition to abnormal adipose tissue metabolism, an increase in simple sugar intake and DNL characterizes obese a compared to nonobese subjects (see later).
Over 50 years ago, Jean Vague classified obese subjects according to the degree of “masculine differentiation”  into those with “gynoid” and those with “android” obesity. Gynoid obesity was characterized by lower-body deposition of fat (around the thighs and buttocks, “pear-shaped”) and relative underdevelopment of the musculature, while android obesity defined upper-body (truncal, “apple-type”) adiposity, greater overall muscular development and a tendency to develop hypertension, diabetes, atherosclerosis, and gout. These phenotypic observations have subsequently been confirmed in many cross-sectional and prospective studies (see  for review).The amount of visceral fat correlates closely with the amount of fat in the liver . Although abdominal obesity is a marker of NAFLD and the MetS, it is controversial whether visceral fat is “causal or correlative”  or “the major culprit or an innocent bystander”  for features of the MetS or NAFLD.The small visceral depot has been suggested to be harmful because omental and mesenteric adipocytes have a higher rate of lipolysis than subcutaneous ones, and this could liberate excessive amounts of FFA into the portal vein and make the liver fatty. Visceral fat may also release more inflammatory cytokines than subcutaneous adipose tissue .
However, this “portal theory” can be questioned . When groups of subjects, who have similar amounts of liver fat but differ with respect to the amount of subcutaneous and visceral fat, are compared, there are no defects in insulin suppression of lipolysis or in rates of VLDL production . If overactive lipolysis were responsible for the harmful effects of visceral fat, one would predict this depot to be reduced rather than enlarged. Surgical removal of omental fat in humans does not improve hepatic insulin sensitivity .
Patients with NAFLD frequently have obstructive sleep apnea (OSA). According to two meta-analyses pooling data from 18  and 11  studies, patients with OSA have a significantly and 2.0–2.6-fold increased risk of having NAFLD and a significantly higher risk of liver fibrosis independent of obesity . This increased risk could be mediated by hypoxia (see  for review). Patients with the polycystic ovary syndrome are more insulin resistant and have a higher liver fat content than equally obese subjects without the syndrome .
Cause of NAFLD: high simple sugar intake and de novo lipogenesis
The increase in the consumption of foods containing high amounts of added simple sugars such as soft-drinks, desserts, cookies, sugar and candy, and fruit drinks has far exceeded that in intake of any other food group . Recent prospective epidemiologic data have shown that an increase in the intake of sugar-sweetened beverages (SSB) predicts both obesity, type 2 diabetes, CVD, and NASH independent of other factors such as energy intake. In the HPFU (Health Professionals Follow-Up Study), the risk of CVD was significantly increased in men using high amounts of SSB after adjusting for confounders such as physical activity and energy intake. Intake of SSB was also significantly associated with features of the MetS such as hypertriglyceridemia, low HDL cholesterol, and high CRP and IL-6 concentrations . There were similar findings in the Nurses’ Health Study. Analysis of food questionnaires from 427 adults with biopsy-proven NASH found fructose consumption, after adjusting for age, gender, BMI and total caloric intake, to be significantly associated with fibrosis . Taken together these data suggest that simple sugars significantly contribute to the epidemics of obesity, type 2 diabetes, CVD, and NASH.
Simple sugars are converted to fat in the liver via DNL . This pathway accounts for some 5% of liver fat synthesis in normal subjects but is threefold higher in patients with NAFLD when compared to equally obese subjects without an increased liver fat content . It produces exclusively saturated fatty acids , which relative proportion in the liver , as well as rate of production increases in direct proportion to liver fat in humans . Saturated fats produced by DNL as compared to fatty acids derived from the diet may cause more lipotoxic injury (see  for review). Some of fructose may be metabolized in the colon and influence gut microbiome . Overfeeding a high carbohydrate (1000 extra simple sugar calories as candies and sugar-sweetened beverages/day) diet for merely 3 weeks increases DNL, liver fat (by 30%), and serum triglycerides and lowers HDL cholesterol . Meta-analyses comparing isocaloric substitution of carbohydrate by fat (saturated or monounsaturated or polyunsaturated) have shown that high carbohydrate but not high fat diets produce the dyslipidemia typical of the MetS and NAFLD . The rate of DNL is increased in obese subjects who are hyperinsulinemic but not in those who are normoinsulinemic, even when obese subjects are consuming the same diet as nonobese subjects . These data imply that excessive intake of simple sugars recapitulates key features of the MetS as well as NAFLD, and that DNL is also increased in hyperinsulinemic subjects irrespective of diet Composition.