Obesity is a physiological state characterized by excessive storage of triglyceride in adipose depots. Because accumulation of triglycerides in adipocytes requires a positive net energy supply, it follows logically that the greatest dietary contribution to obesity would be made by the dietary source or sources that supply the most energy. Therefore, this study was initiated to determine whether fructose could have provided sufficient energy to cause the dramatic rise in obesity since the early 1970s and, if not, what other explanations may exist. To achieve this goal, USDA loss-adjusted food availability data were collected to assess nutrient trends of the average, non-institutionalized American. The phrase “food availability” is often used interchangeably with the phrase “food disappearance,” highlighting the usability of these data as a proxy for food consumption. These data are therefore presented as an estimate of the amount of foods and nutrients consumed daily by the average American. The data do not provide information on specific demographic groups, but rather the entire U.S. population on a per capita basis. The analysis is strengthened by our ability to track trends over several decades using data collected annually. Furthermore, the correction of these data for typical losses such as plate waste and inedible portions allows for more accurate estimates and is a unique characteristic of this analysis as compared to other fructose consumption analyses that use unadjusted USDA data. An additional unique quality of this analysis is the consideration of the amount of individual monosaccharides and disaccharides (whether consumed as simple or complex carbohydrates) by applying the nutrient composition of each food to the mass of each food available for consumption.
The availability of carbohydrates in the US food supply has increased more than any other macronutrient since 1970. In 2009, 32.3 g/d more carbohydrate was available than in 1970, suggesting that carbohydrates alone contributed 129 kcal/d more in 2009 than 1970 with a range of 253 kcal/d. The net accumulation of carbohydrate across the study period as measured by AUC was 9.8%. It is important to ask how much of this increase may be attributed to fructose from all sources. Estimates of total fructose consumption have been reported using the unadjusted USDA food availability data
 and survey data from the Nationwide Food Consumption Surveys, the Continuing Survey of Food Intake by Individuals, and the NHANES
, with the conclusion that per capita fructose consumption increased 18% between 1970 and 2004. The purported increase was suggested as support for a causal role of fructose in obesity. In the present study using loss-adjusted data, we observed fluctuations in fructose availability that included periods of increase and decrease between 1970 and 2009, with an overall result of no net change in total fructose availability.
Several factors must be considered to properly interpret the biological significance of the fluctuations in fructose availability. First, we observed a maximum range of increase in total fructose availability of 11.3 g/d between 1982 and 1999, representing an increase of 45 kcal/d during this 18 y period. It is tempting to associate this increase with the greatest rise in obesity that was observed between 1980 and 2000. However, the increase of 45 kcal/d was minor compared to the increase in total glucose (49.8 g/d) and fat (11.9 g/d) availability that together accounted for an increase of 306 kcal/d during the same 18 y period. While our analysis did show an increase in total fructose availability from 1982 to 1999, emphasizing this change in isolation without considering the much larger changes in glucose, fat and total energy availability unduly magnifies and distorts the contribution of fructose to the rise in obesity. Previous reports that have attempted to link fructose consumption and obesity have not taken into account the comparatively large increases in glucose, fat and total energy availability that have occurred since 1970
[4, 23]. Second, our analysis indicated a decrease in caloric sweeteners and total fructose availability from 1999 to 2009 while obesity trends continued to increase during the same period, illustrating a lack of association between fructose consumption and obesity. Sun et al.
, using the NHANES 1999-2006 databases, also reported no positive association between fructose consumption and body mass index or waist circumference. The increase in fat availability since 1999 appears to account for the continued high level of total energy available. Third, using unadjusted USDA food availability data overestimates nutrient consumption by as much as 30-75% compared to USDA loss-adjusted data, as estimated by the difference between the raw and unadjusted numbers of foods with the greatest and smallest adjustments (data not shown). Our analysis using loss-adjusted data indicated that total fructose availability in 1970 and 2009 were nearly identical; the range of fructose availability was only 11.3 g/d over the 40 y period; and its net accumulation was effectively zero despite significant changes in the food systems that provide fructose, such as using HFCS as a substitute for sucrose in foods and beverages. During the same time, the prevalence of obesity in the US significantly increased, indicating that dietary fructose per se could not have played a quantitatively important role in the increased prevalence of obesity.
Despite the current findings showing a lack of association between total fructose availability and obesity prevalence, one should not regard fructose as a benign nutrient that can be consumed without consequence, particularly if over-consumed. Fructose is a lipogenic nutrient and metabolized differently than glucose
. Relative to glucose, excessive consumption of fructose (25% of total calories) was reported to elevate plasma concentrations of apoB, triglycerides, and LDL cholesterol
. Fructose may also elevate circulating levels of uric acid and multiple liver enzymes relative to glucose
, although a recent meta-analysis of 21 trials failed to show a uric acid-increasing effect of isocaloric fructose intake
. In a study of adult men and women, dietary fructose relative to glucose elevated blood flow to regions of the brain that regulate appetite and also reduced circulating levels of satiety hormones
. These observations demonstrate unique features of dietary fructose that require further study, especially in the context of excess fructose and excess total calories
[8–10, 30–32]. The current study suggests that the unique lipogenic properties of fructose may have been of minor importance to the rise in obesity due to the small contribution of dietary fructose relative to glucose and total energy availability.
The present findings also cast doubt on the purported role of HFCS as a singly important dietary factor in promoting obesity. Despite increased usage of HFCS in the US food supply, no net change in total fructose availability occurred between 1970 and 2009 when analyzed using loss-adjusted data. A critical review of epidemiologic studies and randomized controlled trials failed to demonstrate a relationship between HFCS consumption and increased obesity prevalence
. Since the introduction of HFCS, their usage has been accompanied by decreased usage of sucrose. Unfortunately, no standard methodology or application of HFCS has been employed in human and animal studies, leading to inconsistent results
. Furthermore, the US diet contains products made with both HFCS-55 and HFCS-42, which when consumed in combination within the context of the entire diet, yields nearly identical availability of glucose and fructose. Studies that test only HFCS-55 in comparison to sucrose have a disproportionately higher dietary fructose-to-glucose ratio
If fructose and HFCS are unlikely to have contributed to obesity in a direct manner, then what other dietary nutrients and food sources might be responsible for the increase in total energy between 1970 and 2009? Carbohydrate availability increased more than any other macronutrient. When calculated on a monosaccharide basis from all food sources, the most abundant carbohydrate was glucose. In 1970, there was approximately 3.1-times more glucose than fructose available in the food supply; in 2009, there was 3.6-times more glucose than fructose. The increase in glucose availability from 1970 to 2009--and the lack of an increase in fructose availability--was due to an increased availability of glucose-containing food sources other than caloric sweeteners. Indeed, our findings indicate that the grains category provided more energy and increased more than any other glucose source.
Dietary fats/oils availability also increased from 1970 to 2009. Starting in 1999, fats/oils availability increased sharply while carbohydrate availability began to decline. This apparent replacement of carbohydrate with fats/oils may have been due to increasing popularity of diets lower in carbohydrate and higher in fat, thus accounting for the greater net accumulation of energy from fat versus carbohydrate. Before this macronutrient shift in 1999, carbohydrate had accumulated to a greater extent than fats/oils regarding both mass and energy contribution. Therefore, it appears that from 1970 to 1999, carbohydrate made the greatest contribution to the increased energy availability of the US diet, whereas the fats/oils category was a more important contributor to the energy increase after 1999. The increased availability of the grains and fats/oils categories was further demonstrated by the AUC approach we employed to estimate net accumulation of energy in relation to 1970 food availabilities. This finding is consistent with other food trend studies using both NHANES and the Nationwide Food Consumption Survey in which total energy and total carbohydrate intake have increased since 1970