The rapid increase in the prevalence of obesity most likely reflects changes in lifestyle, such as less activity, less outdoor time, and more sedentary behavior, or changes in nutrition, such as change in diet composition and associated increased intake, which may be exacerbated by genetic predisposition. Increased consumption of sugar and, more specifically, high-fructose corn syrup, is one of the strongest candidates.¹⁶ Findings from large cross-sectional studies in conjunction with those from well-powered prospective cohort studies with long periods of follow-up have shown a positive association between higher rates of soft drink consumption and caloric intake, weight gain, and obesity in both children and adults.¹⁷ Studies also suggest that high consumption of sugar is associated with a higher risk for hypertension and coronary heart diseases,¹⁸ metabolic syndrome and type 2 diabetes,¹⁹ and NAFLD.²⁰ 

Despite the reported deleterious health effects of a high intake of fructose, the Corn Refiners Association has actively campaigned to emphasize that industrially produced high-fructose corn syrup is “natural” and “does not appear to contribute to obesity more than other caloric sweeteners.”²¹ Some meta-analyses claim that the detrimental effect of fructose and other sugars on lipid levels is confounded by the presence of positive energy balance and that without excess calories, sugar consumption does not lead to metabolic abnormalities.²² 

As far back as 2500 BCE, ancient Egyptians deliberately force-fed carbohydrates to birds to fatten them and to produce fatty liver, or foie gras. The biological significance of DNL in mammals remains controversial to this day. In 1860, German chemist Justus von Liebig was the first to postulate the conversion of sugar to fat. In the 1950s, P. Roy Vagelos, MD, and Salih J. Wakil, PhD, demonstrated that DNL was a cytosolic pathway completely distinct from mitochondrial fatty acid β-oxidation.¹⁵ At the time, because of technical limitations in the ability to quantitate DNL, it was believed to occur rarely in humans and to be a physiologically unimportant pathway. 

In the 1990s, Hellerstein et al²³ proposed a revolutionary stable isotope tracer method coupled with mass spectrometry and mathematical modeling that allowed investigators to noninvasively measure DNL in humans. Since that time, studies have demonstrated that high-sugar diets and hepatic DNL can increase the risk for elevated blood lipid levels, cardiovascular disease, NAFLD, insulin resistance, and type 2 diabetes mellitis.²⁴ 

In healthy people consuming a low- or normal-carbohydrate diet, DNL measured in circulating very low-density lipoprotein triglycerides (VLDL-TG) is quantitatively minimal in the fasting state (<3% of VLDL-TG), but it increases approximately 10-fold in the postprandial state (20%-30% of VLDL-TG) because of the rise in lipogenic precursors that occurs after the consumption of a meal.²⁵ In contrast, data have shown that fasting DNL is quantitatively statistically significant under the following conditions: (1) after 5 days of energy-balanced, high-carbohydrate feeding (65%-70% of calories; Figure 1) and (2) after 5 days of overfeeding with carbohydrate or fructose in excess of energy (Figure 2).^26,27 

These data demonstrate that fractional DNL is not detectable during caloric restriction but increases progressively to 20% in the fasting state and 30% to 35% in the fed state after 5 days of carbohydrate overfeeding. These studies also show that DNL is a very sensitive and specific marker of recent carbohydrate intake and energy balance and highlight the importance of considering the state of energy balance, in addition to carbohydrate intake, in studies of DNL. 

In addition, these studies established that both isoenergetic and hypercaloric diets high in carbohydrates, especially fructose, are highly lipogenic and produce elevations in DNL that persist after an overnight fast. It should be noted, however, that carbohydrate-induced DNL appears to be specific to simple carbohydrates. Timlin and Parks²⁵ demonstrated that when simple sugars are restricted and a diet high in complex carbohydrates (eg, starch) is consumed, DNL is trivial after an overnight fast. 

Likewise, in the fed state, DNL is also determined by the type of simple sugar consumed. Fructose, but not glucose, increased hepatic DNL in 6 healthy lean participants (Figure 3).⁹ During 6 hours of fructose ingestion, DNL increased 20-fold, and 25% of circulating VLDL-TG was derived from DNL. In contrast, when the study was repeated in the same participants using glucose levels, rates of DNL were unaffected, and only 1% to 2% of VLDL-TG was synthesized de novo. These data demonstrate that fructose is a potent stimulus to lipogenesis. 

Fructose metabolism is unique because, in contrast to glucose, it is metabolized almost exclusively in the liver, as its initial metabolism is hepatic. In contrast, glucose is predominantly metabolized in extrahepatic tissues, with the liver accounting for only 20% of glucose metabolism.¹⁴ Fructose is metabolized more rapidly than glucose, providing precursors, such as acetyl coenzyme A, for DNL. Animal studies have shown that fructose is converted to fatty acids at rates of up to 18.9 times faster than glucose¹⁴ and produces increased TG levels²⁸ and liver fat.²⁹ Overall, fructose is a potent lipogenic substrate in humans, and it significantly increases plasma TG levels.^10,11 

There is evidence that both hyperinsulinemia and a diet rich in simple sugars, by promoting DNL, contribute to hypertriglyceridemia and liver fat accumulation and, ultimately, to metabolic syndrome.³⁰ 

A large proportion of people with obesity or NAFLD have insulin resistance,³¹ and most people with insulin resistance have increased liver fat concentration.³² Fabbrini et al³³ demonstrated that insulin resistance and other metabolic perturbations in obesity correlate better with liver fat than with visceral adipose tissue. In contrast to observations in healthy individuals, participants with insulin resistance and NAFLD have been found to have elevated rates of DNL in both the fasting and postprandial states (26% of VLDL-TG).³⁴ Moreover, analysis of liver biopsy specimens collected from participants with NAFLD and chronically elevated DNL revealed that 26% of their liver fat was produced de novo,³⁴ supporting the hypothesis that hepatic DNL may be an important contributor not only to postprandial lipids secreted by the liver but also to liver fat accumulation. 

Figure 4 illustrates the relationship between hyperinsulinemia and DNL in the following conditions of hyperinsulinemia, in which fractional DNL is higher: 

A highly controlled metabolic ward study¹² was conducted to test the hypothesis that short-term increases in hepatic DNL would be accompanied by detectable increases in liver fat concentration. The effects of a fructose-rich diet (25% of kilocalories) were compared with an isoenergetic diet containing the same macronutrient distribution but with low fructose (<4% of kilocalories) on liver fat concentration and hepatic DNL in healthy participants.¹² Eight healthy men were studied as inpatients for 18 days to control energy balance and assure strict adherence to the diet, which was formulated and prepared by bionutritionists. The diets were given during consecutive 9-day periods. Four of the participants consumed the low-fructose diet first, and the other 4 started with the high-fructose diet. Fractional hepatic DNL during feeding was measured using stable isotope tracers and mass isotopomer distribution analyses. Liver fat concentration was measured on day 8 of each dietary period by magnetic resonance spectrometry on a 3-Tesla scanner.¹² 

Given the aforementioned evidence that DNL can be acutely increased with hypercaloric feeding and decreased during hypocaloric feeding, weight was monitored daily and diets adjusted to carefully maintain a neutral energy balance. As a result, there were no net changes in weight or whole-body composition, measured by dual-energy x-ray absorptiometry, during the study. After 8 days consuming the high-fructose diet, mean (SD) DNL during feeding was significantly higher with the low-fructose group (18.6% [1.4%] vs 11.0% [1.4%]; P=.001). Liver fat concentration was also higher during high-fructose feeding (median, 137% of values obtained during the low-fructose diet; P=.013). These differences were observed regardless of the order in which the diets were consumed.¹² Hepatic glucose production, which is typically suppressed by high insulin levels, was not blunted by a high-fructose diet. This hepatic insulin resistance may contribute to the development of type 2 diabetes. Hepatic glucose control can be restored by restricting added sucrose and fructose in one's diet. 

Taken together, these results demonstrate that after as few as 8 days and in the presence of a neutral energy balance, a high fructose intake can increase both hepatic DNL and liver fat concentration. Perhaps most importantly, this study demonstrated that even short-term fructose restriction is associated with demonstrable reductions in DNL and liver fat concentration and provided the basis for a subsequent study of fructose restriction in children. 

In an outpatient study¹³ in obese Latino and African American children with features of metabolic syndrome (eg, elevated triglycerides, insulin resistance, hypertension) who reported high levels of sugar intake, 9 days of fructose restriction with isocaloric substitution of complex carbohydrates resulted in reductions in fasting glucose and insulin levels, improved glucose tolerance,⁴¹ and less atherogenic lipoprotein profiles.⁴² Importantly, these same participants had a 56% decrease in DNL during feeding (n=40).¹³ Although the children in this study were not selected for the presence of liver fat concentration, approximately two-thirds of them had elevated liver fat concentration (>5% liver fat fraction by magnetic resonance spectroscopy). Overall, liver fat concentration decreased by 22% during fructose restriction.¹³ This study provides further evidence that carbohydrate quality, specifically fructose, plays an important role in liver fat accumulation and hypertriglyceridemia and that DNL is an important factor in these processes. 

For decades, the beverage industry and investigators sponsored by the industry have continued to mount vigorous and well-funded campaigns against public health policies intended to encourage limitation of soda and sugar consumption. Recent reports document that many institutions and respected academicians have received support from the beverage and sugar industry⁴³ and have described the extensive efforts of industry groups to maintain control over the messages regarding sugar consumption and metabolic diseases.^6,44 A 2016 publication⁸ strongly urges that policy-making committees should be cognizant of the bias that might be introduced by corporate sponsorship. 

Efforts to reverse the obesity epidemic in the United States have been largely unsuccessful. While the rate of obesity continues to increase globally, there is a growing body of literature demonstrating how lifestyle and dietary interventions can have a profound effect. Many obese people seem to be mired in the fact that the management of obesity is long and very difficult. The present review details data demonstrating that changing the dietary intake of fructose can significantly reduce DNL and its unhealthy metabolic sequelae in as little as 9 days and, furthermore, this benefit can occur even without any loss of weight. This finding allows us, as health care professionals, to reframe the discussion from weight loss to attaining better health in just a week and a half. To quote Andrew Taylor Still, MD, DO,⁴⁵ “To find health should be the object of the doctor. Anyone can find disease.” These studies can be used to help explain to our patients the rationale as to why they should follow the most recent dietary recommendations to limit simple sugars to no more than 5% to 10% of their daily intake. Should patients adhere to these recommendations, it may well be the first step to facilitate their body to self-regulate, self-heal, and begin its return to a state of health maintenance. 

There is increasing evidence from both epidemiologic and interventional studies of the deleterious effects of consuming sugar-sweetened beverages and other foods containing fructose and other added sugars, even in the absence of weight gain. Chronically elevated DNL during both fasting and feeding may be a key mechanism in the link between consumption of added sugar and metabolic abnormalities, including obesity, high lipid levels, cardiovascular disease, NAFLD, insulin resistance, and type 2 diabetes. 

We thank Moises Velasco-Alin, BS, for valuable assistance with the preparation of this manuscript.