Literature Review: Fructose-Sweetened vs Glucose-Sweetened Beverages
So tonight as my husband and I were working on my website, we found that the number one google search click was my assessment of a research article about how fructose-sweetened vs glucose-sweetened beverages affect the body.
The assessment I conducted was based on a critical review form created by an Evidence Based Research Group out of McMaster University in Hamilton, Ontario. Last summer when I started really assessing research studies, I was given permission by this group to use their forms as well as guidelines here on my website. Click here for guidelines on how to assess a quantitative study, and here for guidelines on how to assess a qualitative study. Click here for the quantitative study assessment form, and here for the qualitative assessment form.
While I had planned to just put all my article assessments on the website as files to access, I'm now going to take it one step further and talk about them with you. Each article will have its own post, but be set up in the same format, as you will see below. :)
So, without further delay, let's chat about this article.
Citation:
Reference: Stanhope KL, Schwarz JM, Keim NL, Griffen SC, Bremer AA, Graham JL, et al. 2009. Consuming Fructose-sweetened, not glucose-sweetened beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans. Journal of Clinical Investigation, 119(5): 1322-1334.
Study Purpose:
The purposes were clearly stated: so assess the following questions.
a) Does consumption of fructose with an ad libitum diet promote greater weight gain and have differential effects on regional adipose deposition and adipose gene expression compared with consumption of glucose with an ad libitum diet?
b) Does consumption of fructose induce dyslipidemia compared with consumption of glucose?
c) Is fructose induced hypertriglyceridemia the result of increased rates of hepatic de novo genesis (liver fat) and or decreased triglyceride (TG) clearnance?
d) Are there differences between responses of older men and post-menopausal women to dietary fructose?
Literature Review:
I'm undecided if the literature was reviewed well enough. The authors only sited 4 articles on the effects of fructose consumption, whereas they could have sited more as there were many more cited in the discussion section of the study.
Study Design:
The study design was a randomized control trial (RCT). Double blinded parallel arm using matched subjects over the course of 3 phases.
a) 2 week in-patient baseline – energy balanced diet
b) 8-week outpatient intervention – consumed either fructose or glucose sweetened bev totaling 25% of total energy daily on top of usual diet.
c) 2 week inpatient intervention – energy balanced + 25% total calories of sweetened bev on top of diet.
Biases of the study design:
a) Sample bias – volunteers (through newspaper ads) – usually different people than non-volunteers.
b) Measurement bias:
possibly too many measures for small sample size.
c) Intervention bias:
Contamination – none. All received same treatment
Co-intervention – none. All received same treatment.
Site: all done at same place – UCD Clinical and Translational Science Center’s Clinical research center (CCRC)
d) All seen by same dietitian to ensure same treatment.
Study Sample:
n=39. (n=number of participants in the study), possibly adequate number of participants; however the sample size justification was not provided.
Age: 40-72 yrs, bmi btw 25-35 kg/m2, stable reported weight for last 6 months.
Exclusion – disease states, medications for diabetes and heart disease, high TG, drinking excessive ETOH or extra sugar drinks.
N=39, 7 drop outs (3 glucose group, 4 fructose group) due to inability or willingness to comply.
Fructose group n=17; Glocose group n=15.
Ethics:
Outcomes:
Outcome measures were reliable.
Outcome measures were valid.
Measures:
Daily weight (morning before breakfast)
Total fat (dual energy x-ray absorptiometry (DXA))
Visceral fat (CT scan)
Blood pressure (cuffs) 2x/d during inpatient weeks
BS: 3 fasting and 33 post-prandial blood samples (30 min intervals).
Hepatic fractional de novo lipogenesis (baseline and 10 weeks)
Oral Glucose Tolerance Test (OGTT) after overnight fast. Pre, then 6 post-prandial samples (30 min intervals). 75 g oral glucose load.
Intervention:
Intervention was described in detail.
Contamination of the intervention was avoided by using the same dietitian and providing similar teachings to both groups.
Co-intervnetion of the intervention was avoided.
Intervention description: (could be replicated based on information provided).
Energy balance: 15% protein, 30% fat, 45% carbohydrate (complex carb). Energy calculated with Mifflin equation, 1.3 activity factor at home,1.5 for inpatient days.
In patient diet calorie allocation: 25% breakfast, 35% lunch, 40% lunch. Must eat all, and nothing more (1st 2 weeks); must eat all + sweetened bev (last 2 weeks) – (30% complex carb + 25% sugar sweetened bev).
NOTE: 25% of total calories = ~ 610 kcal/d from just fructose or glucose which is very high. Average intake of fructose in USA = ~ 15% of total calories (ref 59-63). 16% of USA reportedly eating 25%+ total calories from sugar-sweetened bev (ref 59,62,63)
Food intake measures over 8 week out pt ad lib diet + 3 sweetened bev (in koolaid): 24 hour recall (telephone) using USDA 5 step multiple-pass method (Conway et al – ref 69). Done at 2 and 7 weeks. Analyzed with Nutrition Data System for Research.
Compliance measured based on urine samples (bev containing biomarkers (riboflavin), tested 2x per week (at bev pick up)
24-hr blood tests: meals prior to blood work – identical at 2,8, 10 weeks. And on wt maintaining diet for 10 days.
Buffet meals for ad lib days – measured intake without volunteer knowing.
Results:
Results were reported in terms of statistical significance. Analysis methodology was appropriate.
Baseline results: no significant differences over first 2 weeks between groups.
8 week out patient intervention: comparably significant wt gain (P<0.01 for glucose and p<0.001 for fructose).
Glucose group – significantly increased in subcutaneous fat (P<0.01)
Fructose group – significantly increased in total abdominal fat (P<0.01)
and visceral adipose tissue (area around the belly) (p<0.001).
Gene expression (based on lipogenic and other genes) changed more for glucose than fructose (for stearoyl-CoA desaturase, fatty acid desaturase 1, and fatty acid desaturase 2).
Blood Pressure was normal for all clients at baseline. No changes over intervention at 10 weeks. [however, it is important to note that BP was not reported to be taken over the 8 week intervention, just at 9 weeks when they were at the facility for 2 weeks].
Fasting TG concentrations: increased in glucose group (P<0.01); marked variability change in fructose group (indices for post-prandial TGs increased in fructose group (p<0.001)); change in area under the curve increased in fructose group (p<0.0001)
Fasting LDL number – increased in fructose group (P<0.01), not glucose group
Fasting small-dense LDL (athrosclorotic LDL) increased in both the fructose group (p<0.001) and the glucose group (p<0.01)
Fasting oxidized LDL increased in the fructose group (p<0.01)
Fasting ApoB (p<0.001), post prandial ApoB (p<0.001) and the ApoB:ApoA1 ratio (p<0.001) increased in the fructose group
Total cholesterol number increased in the fructose group (p<0.0001)
Post-prandial concentrations of plasma remnant-like particle lipoprotein –cholesterol (RLP-C) (p<0.0001) and remnant-like particle lipoprotein – triglycerides (RLP-TG) (p<0.0001) increased in fructose group.
Post prandial de novo lipogenesis was significantly increased in fructose group (P=0.021)
Post prandial post-heparin lipoprotein lipase (LPL) activity tended to decrease after 10 weeks in fructose group; where as in the glucose group it tended to increase. [possibly meaning reduced TG clearance?] but this was a trend, not significant… so cannot make any conclusions from it.
Overall difference in the glucose vs fructose group was significant: p=0.041)
Insulin sensitivity decreased (participants became more insulin resistant) in the fructose group – increased fasting insulin levels in the fructose group; and slower glucose clearance at 30 (p<0.05), 60 (p<0.01), 90 (p<0.001), 120 (p<0.001), and 180 (p<0.05) minutes post prandial.
Clinical Importance was reported as well.
It appears that very excessive glucose, although it leads to weight gain (subcutaneous fat) and small density LDL (atherosclerotic LDL), does not seem to increase the other markers for metabolic syndrome and cardiovascular disease.
Drop outs were reported (see study sample section). Dropouts = 7 (3 from glucose and 4 from fructose groups). Reasons listed was non-compliance, and inability to continue with study.
Conclusions and Implications:
Conclusions were likely appropriate given the study methods and results (although possibly low sample size given the number of outcome measures..
Fructose on its own, in very high doses increases many factors which can contribute to heart disease: TGs, Total cholesterol, total LDL number, small density LDL number, number of oxidized LDL, insulin resistance (fasting plasma glucose and fasting insulin), visceral fat weight gain (around the middle), apoB and apoB:apoA1 ratio.
Although at high levels glucose still causes weight gain and can increase risk of CVD (elevated number of small-dense LDL (atherosclerotic particles)), it affects far fewer risk factors for chronic disease than fructose does.
Bottom line, both Fructose and Glucose need to be consumed in moderation. Trying not to drink sugar is a good way of being able to listen to your body’s fullness cues (as drinking sugar ADDED to caloric intake), and when you choose to eat sugar/fructose/glucose, eat it mindfully.
More studies are required to test the effects of fructose in lower consumption (closer levels to the population).
Limitations:
Smaller study, but did look at pre/post levels from the same person, which makes the evidence a bit stronger.
Sampling method – volunteers – different people (with different daily habits and routines than people who would not normally volunteer
Possibly too many measure for this study with the sample size.
(Please note, i'm in the process of adding a number of posts with references, pictures, etc. Rather than wait until every thing is perfect, I'm adding it now!)