It has been suggested that the consumption of fructose, which is consumed mainly in the form of HFCS and sucrose, may be a contributing factor to the increased incidence of obesity and metabolic syndrome (1
). The current estimate for the mean intake of added sugars by Americans is 15.8% of energy. However this value is based on consumption data from the 1994-1996 Continuing Survey of Food Intakes by Individuals (CSFII) (3
), and recent reports suggest that energy intake from sugar-sweetened beverages alone approaches or exceeds 15% of calories in several population groups (4
). The large standard deviations in several of these reports suggest that at least 16% of the studied populations were consuming greater than two times the mean intake, and likely well over 25% of daily energy requirements from sugar-sweetened beverages (5
). Based on these reports it is reasonable to estimate that the percentage of energy consumed as sugar from both beverages and solid food is greater than 20% in a significant portion of the U.S. population.
We have reported that consuming fructose-sweetened beverages with meals results in lower 24-h circulating glucose, insulin and leptin concentrations, and decreased postprandial suppression of plasma ghrelin levels when compared with consumption of glucose-sweetened beverages (9
). We and others have reported that consuming fructose-sweetened beverages increases postprandial triacylglycerol (TG) concentration compared with glucose-sweetened beverage (9
), and that these responses are more pronounced in men compared with women (10
), and in overweight/obese subjects compared with normal weight subjects (11
). Since insulin, leptin, and possibly ghrelin function as key signals to the CNS in the long-term regulation of energy balance, prolonged consumption of diets high in energy from fructose could lead to increased caloric intake and contribute to weight gain and obesity (2
). The sustained elevation of plasma TG concentrations after fructose ingestion suggests that chronic over-consumption of fructose could also contribute to atherogenesis and cardiovascular disease (12
). However, pure fructose and pure glucose are not commonly employed as sweeteners. Until a few decades ago, most foods and beverages in the U.S. were sweetened with the disaccharide sucrose, which is composed of 50% glucose and 50% fructose. In 1970 an enzymatic process to convert corn sugar (composed of glucose) into high fructose corn syrup (HFCS) was developed. Since then, HFCS, mainly in the form containing 55% fructose and 45% glucose (HFCS-55), has replaced sucrose as the predominant sweetener used in soft drinks and represents approximately 40% of sweeteners added to foods consumed in the U.S. (14
There are currently few studies comparing the metabolic and endocrine effects of HFCS with sucrose (15
), and to our knowledge, there are no studies comparing the effects of consuming HFCS with pure fructose or glucose. Therefore the objectives of this study were to compare the metabolic and endocrine effects of consuming HFCS- and sucrose-sweetened beverages, and to determine if responses are affected by gender and adiposity. A third objective was to compare the effects of consuming HFCS- and sucrose-sweetened beverages with the consumption of beverages sweetened with fructose or glucose.
Circulating glucose, insulin, leptin, ghrelin, triacylglycerol (TG), and free fatty acid (FFA) concentrations were measured in 34 subjects over a 24-h period on 2 separate days during which they consumed three isocaloric, mixed nutrient meals accompanied by either sucrose-sweetened or HFCS-sweetened beverages. In a subset of 8 male subjects, these variables were also measured on 2 additional days during which these subjects consumed the same isocaloric meals accompanied by beverages sweetened with 100% fructose or 100% glucose.
Thirty-four subjects (18 men and 16 women) with an age range of 20–50 yr (Mean: 34.7 ± 1.7y) participated in the study. Participants were recruited through newspaper advertisements and underwent a telephone interview, a complete blood count, and a serum biochemistry panel to assess eligibility. Respondents with anemia, hepatic or renal disease, diabetes mellitus, fasting serum TG levels >400mg/dL, hypertension, eating disorders, or who had surgery for weight loss were excluded from the study. Individuals who smoked, or who took thyroid, lipid-lowering, glucose-lowering, anti-hypertensive, anti-depressant, or weight loss medications, or who were pregnant or lactating were also excluded from participating. The Institutional Review Board of University of California, Davis approved the experimental protocol, and subjects provided informed consent to participate in the study.
Each subject participated in two experimental trials conducted in random order. The experimental days were spaced 1 month apart, and each required an overnight stay in the University of California, General Clinical Research Center (GCRC). During each experimental day the subjects consumed identical meals based on calculated energy requirements (as described below) that included beverages sweetened with either HFCS or sucrose. The 18 male subjects were invited to extend participation in the study, and of these, a subset of 8 men completed 2 additional study days during which they consumed the identical meals accompanied by beverages sweetened with either 100% fructose or glucose. Participation in the additional fructose and glucose trials was limited to male subjects due to budgetary constraints.
Percent body fat was determined by measurement of bioelectrical impedance (BIA) (Tanita 310GS Body Composition Analyzer, Tanita Corp, Tokyo, Japan). BIA measurements of body fat in healthy adults has been shown to correlate well with body fat measurements by duel-energy X-ray absoptiometry (DXA) (17
). As there is no consensus on the percent body fat standards for overweight and obesity (19
), we used a statistical approach of dividing subjects of each gender into two equal groups based on ranking by percent body fat. Therefore, the men and women were divided into two groups with lower and higher body adiposity based on a percent body fat lesser or greater than 22% and 32%, respectively.
The subjects were instructed to maintain their normal dietary intake and level of physical activity during the interval between the GCRC studies. Following a 12-h overnight fast, the subjects checked into the GCRC at 0700 h. A physical examination was conducted by the study physician, and an intravenous catheter was inserted and kept patent with a slow saline infusion. Blood sampling commenced at 0800 h and continued for 24 hours. Thirty-six blood samples were collected for substrate and hormone measurements over the 24-h sampling period during which the subjects consumed three standardized meals. Each meal was accompanied by a sucrose-, HFCS-, glucose-, or fructose-sweetened beverage.
The meals consisted of whole foods and were designed by a registered dietitian. The nutrient composition of the diets was determined using the Food Processor SQL, ESHA Research, Inc. (Salem, OR). Each subject ingested three meals per day. Breakfast consisted of scrambled eggs, ham, potatoes, and asparagus. Lunch consisted of chicken/tortilla soup with corn chips and cheese. Dinner consisted of lasagna with beef and a salad with lettuce, tomato, cucumber, celery, cheese and vinaigrette dressing. The energy content of the meals was based on each subject's daily energy requirement as estimated by the Mifflin equation with an activity factor of 1.3 (20
). The activity factor was low because the subjects remained relatively sedentary while at the GCRC. Twenty percent of the energy requirement was consumed at breakfast, 35% at lunch, and 45% at dinner. The meals contained 30% energy from fat, 15% from protein, 30% from complex carbohydrate and 25% of energy from sugar. The 25% sugar consisted of sucrose, HFCS, glucose, or fructose in the form of a beverage. It is important to study the effects of sugars at the 25% of energy intake level. The Institute of Medicine of the National Academies in the 2002 Dietary References Intakes concluded that there was insufficient evidence to set an upper intake level for added sugars since there were not specific adverse health outcomes associated with excessive intake (21
). Therefore, they suggested a maximal intake level of 25% of energy intake from added sugars. The sucrose and HFCS beverages were provided by the sponsor (PepsiCo, Inc., Purchase, NY) as 11% w/w sugar in non-caffeinated, carbonated sodas. The fructose and glucose were prepared as 11% w/w solutions in carbonated soda water, flavored with a commercial unsweetened drink mix. The subjects and GCRC study staff were blinded to the sweetener contained in the beverages. The beverages were not matched for sweetness, but as the trials were spaced approximately one month apart it not likely the subjects would have noticed a difference in the sweetness of the beverages. The only other beverage served during the 24-h study period was water. Breakfast was consumed at 0900 h, lunch at 1300 h, and dinner at 1800 h. The subjects were required to ingest all of the provided food and beverage within 20 minutes, and were observed to ensure compliance.
Blood samples were drawn from the catheters at 30-min intervals around periods of meal ingestion and during the predicted nocturnal rise of plasma leptin concentrations (i.e.
5 hours after the evening meal) and at hourly intervals at other times (22
). After the three baseline samples, which were collected at 0800, 0830, and 0900 hours before ingestion of the first meal, a total of 33 additional samples were collected 30–60 min apart. Each sample collection involved the removal of 1 ml of blood to clear the catheter tubing, followed by a 5mL collection into blood collection tubes containing EDTA. Samples were then centrifuged, aliquoted, and stored at -80 °C until assayed.
Assays and data analysis
Plasma glucose and lactate concentrations were measured with an automatic analyzer (YSI 2300 StatPlus Glucose Analyzer, Yellow Springs Instruments, Yellow Springs, OH). Insulin and leptin concentration were measured by radioimmunoassay (RIA) (Linco Research, Inc., St Charles, MO). Total ghrelin concentrations were measured in unextracted plasma with an RIA (Phoenix Peptide, Phoenix, AZ). TG concentrations were measured with an automatic analyzer using maunufacturer's reagents (PolyChem Analyzer, PolyMedCo, Inc., Cortlandt Manor, NY). FFAs were measured with enzymatic colormetric reagents (Wako Chemicals, Richmond, VA). Fasting lipid concentrations were measured at baseline and at 0800h the following morning, and apolipoprotein B100 (ApoB) concentrations were measured at baseline and at 2200h with an automatic analyzer (PolyChem Analyzer).
The area under the curve (AUC) was calculated for glucose, insulin, ghrelin, FFA and TG with a data spreadsheet program (Microsoft Excel; Microsoft, Redmond, WA) using the trapezoidal method. The mean of the three baseline values was determined, and net AUC was calculated by subtracting the areas below baseline from AUC values above baseline. AUCs for glucose, insulin, ghrelin, FFA and TG are expressed as units per 23 hours above each subject's fasting baseline levels because the three samples during the first hour determined the baseline levels. The nadirs for plasma leptin were determined as the two lowest consecutive morning values before 1200 h as previously described (22
). The AUCs for leptin are therefore expressed as units above each subject's nadir over 24 h. Data from the 34 subjects were analyzed with the statistical software programs SPSS 15.0 (SPSS Inc. Chicago, IL) using the general linear model for repeated measures that included a sugar × adiposity group × gender interaction. Age was also included in the analysis with subjects grouped by 20-29, 30-39 and 40-50 years. Time was included as a within subject factor to assess change from fasting levels in lipid concentrations. When the general linear model results indicated the presence of a significant sugar effect, or gender or body adiposity interaction, paired and unpaired t tests were performed within group, or between men and women or between adiposity groups. Differences between the responses to the four sugars in the subset of 8 male subjects were assessed with GraphPad Prism v.4.03 (San Diego, CA, USA) using repeated measures 1-factor ANOVA, and post-tests were performed using Tukey's test for multiple comparisons.