As I was about one-third of my way through writing about this study, I realized that I was biased. My understanding of the literature surrounding fructose has led me to conclude that most studies showing detrimental effects are not applicable to the general population (let alone a healthy population) because of methodical limitations such as using fructose in isolation, using doses well above reasonable consumption levels, and using rodents that have known differences from humans in fructose metabolism (e.g., the fact that de novo lipogenesis under conditions of long-term, high-carbohydrate feeding accounts for 60% to 70% of fatty acids in rodents but less than 5% in humans).
I therefore deleted everything I had and replaced it with what you are now reading. My goal this time around is to be more objective. As such, I apologize in advance to those of you who find my focus on the nitty-gritty annoying, but I promise to also present a practical conclusion.
I feel that this extra scrutiny is further required by the fact that the authors state no conflict of interest, yet 3 of the 4 authors work at the Rippe Lifestyle Institute in Florida, which has received funding from ConAgra Foods, PepsiCo International, Kraft Foods, Coca Cola, Pepper/Snapple Group, the Corn Refiners Association, Weight Watchers International and Kellogg’s. The study at hand was also supported by a grant from the Corn Refiners Association. So yah, clearly no conflict of interest.
I realize that you still have no idea what I’m talking about aside from the slightly alluding article title. The study and topic of this article was a recent publication by Angelopoulos et al, who sought to evaluate the effects of consuming a realistic amount of fructose in its various forms on components of metabolic syndrome among healthy adults over a 10-week intervention period.
A total of 366 healthy, middle-aged men and women with a BMI between 23 and 35 (average 26.3) were recruited and randomized to consume low-fat milk that had been sweetened with 9% of the participant’s estimated energy requirements from fructose (the average intake in the US), 9% from glucose, 18% from sucrose, or 18% from high-fructose corn syrup (HFCS). The higher percentage of calories from sucrose and HFCS was to equilibrate the fructose content, since they contain half fructose and half glucose.
Milk was provided weekly during individual counseling sessions so as to prevent discussion between participants about the amount of milk they received and its sweetness level. Participants received between 2 and 6.5 eight-ounce containers of milk per day depending on their calorie requirements, each containing either 15 grams of added sugar (fructose or glucose condition) or 30 grams of sugar (HFCS or sucrose conditions). Accordingly, someone on a 2000 kcal diet would be consuming 45 grams of fructose or glucose, or 90 grams of sucrose or HFCS daily.
The study was double-blinded to the researchers and participants, but the Rippe Institute staff were aware of the sugar content (but not type) of the milk beverages so as to ensure they provided the correct ones during the counseling sessions. Each participant was provided $300 compensation for completing the 10-week intervention. Measurements were made before and after.
Of the 366 recruited participants, only 267 completed the intervention. This 27% drop-out rate was below the 33% drop-out the researchers had planned for when determining how many participants they would need to maintain sufficient statistical power.
An interesting note about the drop-outs is that the majority were listed as “non-compliant” with no further information provided. This indicates that the participants were not drinking their prescribed milks, but leaves the reasoning open to imagination. Could it be that these individuals noticed unfavorable changes in weight or health and therefore decided to stop drinking the milk? It is a possibility, and one that becomes more concerning when we consider that only the completer data was used for analysis. Accordingly, the results may be underestimated.
Arguably the most important information in dietary interventions is what actually happened with the participant’s diet. Before and after the 10-week intervention, each participant filled out a 3-day food log. Ignoring the potential inaccuracies of food logs (usually under-reporting), caloric intake, sugar, and the percentage of calories supplied by carbohydrates significantly increased vs baseline when all groups were analyzed together. This was certainly to be expected given the intervention.
When the groups were compared to one-another, the change in sugar consumption and percentage of calories supplied by carbohydrates increased to a significantly greater extent in the sucrose and HFCS groups compared to the glucose and fructose groups, which was again to be expected considering they consumed twice as much sugar via the dietary intervention.
Calories supplied by fat was significantly reduced among all groups combined, with no between-group differences. Protein intake as a percentage of energy remained unchanged, probably because of the milk’s protein content. The change in caloric intake was not significantly different between groups, which is somewhat surprising considering that the HFCS group showed a 24% increase vs 6% in the fructose group and 10% in the glucose and sucrose groups.
Although the diet logs and nutrition data based on them may be inaccurate, the scale weight doesn’t lie. All groups combined gained a significant 2 lbs through the intervention, clearly indicating the presence of a caloric surplus. Theoretically this would require a 100 kcal surplus each day, which is incredibly small compared to the 180 or 360 kcal of sugar plus whatever was provided by milk that someone on a 2000 kcal diet would have been adding into their diet.
As such, compensation did occur, but whether this was owed to the sugar, the milk, or their combination remains unknown. Previous research has shown that protein is the most satiating macronutrient, which the milk so graciously provided. However, previous research using real soda has shown that fructose, glucose, and HFCS have similar effects on satiety and do compensate for about 68% of the provided calories. Accordingly, the current findings are in-line with past research using regular soda instead of sweetened milk.
In-line with the increased body weight, waist circumference was significantly increased in all groups by just under 1 cm. That is as far as the anthropometric measurements go, however, leaving us with no indication about changes to body composition.
Cardiometabolic risk factors
Looking at the groups as a whole, there were significant increases in triglycerides and total cholesterol, no significant changes in LDL-c, HDL-c, fasting glucose, or CRP, and significant reductions in systolic and diastolic blood pressure.
As shown in figure 2, the only significant difference between the groups was in triglycerides, with the HFCS group showing significantly greater increase than the glucose group. The reason for this finding is not known. It is difficult to draw conclusions from the data, as no consistent patterns become apparent. The significant increase in both triglycerides and total cholesterol appears to have been driven by the HFCS and sucrose groups, whereas CRP appeared to be driven up by the glucose and fructose groups.
Cardiometabolic effect of sugar is small in reasonable amounts
If you tell healthy middle-aged adults to start drinking low-fat milk that has been sweetened with added sugar, regardless of the type, they may gain weight through a failure to fully compensate for the additional calories. However, the amount of weight gain is small (~1% bodyweight over 10 weeks) and appears to be of little clinical relevance as most components of metabolic syndrome were not significantly altered, showed a significant but negligible increase, or showed a favorable decrease.
The only factor that demonstrated a notable change was the 10% increase in triglycerides. However, whether this was owed to the sugars per se or the excess consumption of calories remains unknown. A recent meta-analysis regarding fructose’s effect on blood lipids determined that fructose increased triglycerides only when provided under hypercaloric conditions. Accordingly, the increase in the present study may very well have been owed to the caloric surplus.
The possibility remains that the low-fat milk could have off-set the detrimental effects of sugar consumption. One study in obese women noted that consuming 3 servings of low-fat milk per day over a 6-week period significantly reduced HDL-c, fasting glucose, and blood pressure. Similarly, another intervention showed gender-specific changes with low-fat dairy consumption whereby men reduced fasting glucose while women reduced waist circumference and bodyweight. With no milk-only control group to compare to in the current study, we can only speculate.
This study population was also relatively healthy, with cardiometabolic risk factors being within normal ranges and an average BMI that was barely overweight. Would similar effects be observed in less healthy individuals, which happens to be the majority of the general public? If you are relatively healthy, then consuming the population average for fructose intake does not appear to be detrimental. That said, there are clear study design flaws that need to be addressed in future research.