Friday, January 3, 2014

Fructose - An Update: "Fructose Has Adverse Effects Only Insofar as It Contributes to Excess Calories" Plus: The Role of Exercise + Meta-Analyses on BP, Weight Gain & T2D

It's incredible... for some, but probably not unexpected for most of us that fructose becomes problematic in overfeeding scenarios, only.
Some of you will probably have seen the press release from the St. Michel's Hospital that made it onto all the major science outlets on the Internet and up on Alex' Facebook page, where he tagged me and thus got me interested in a study that claims to provide evidence that: "Fructose does not impact emerging indicator for cardiovascular disease" | read more.

The main goal of the corresponding paper that has been published in the Atherosclerosis earlier this months (Wang. 2014) was to identify and analyze all clinical interventions that investigated the chronic effect of exchanging isocaloric or hypercaloric oral fructose for a reference carbohydrate on postprandial triglycerides.
Update - Coca Cola & Co buy a white slate for their sugar-sweetened beverages (SSB): Shortly after publishing my analysis of the meta-analysis, I hit onto a more recent review that deals with sugar sweetened beverages and the influence sponsors from the industry have on the outcome of corresponding studies and, more importantly (since easier to be biased) reviews. While the main finding of Bes-Rastrollo's et al.'s analysis is that there is a 5x higher likelihood of SSBs being portrayed as benign, when reviews are financed by the industry, the editor of PLoS|One Medicine rightly points out that "[a] major limitation of the study at hand is however that it cannot assess which interpretation of the available evidence is truly accurate" and that "scientists involved in the systematic reviews that reported having no conflict of interest may have had preexisting prejudices that affected their interpretation of their findings". In other words, financed and non-financed research are both biased" (Editorial published with Bes-Rastrollo. 2013 | learn more).
Don't forget: Nobel Laureate Peter Higgs worked with the method on the right: Conclusion first: "There is a boson that mediates gravitational forces" ➲ Years of research: "Heureka!"
It's also important that you realize that meta-analysis such as the one at hand are less prone to bias, than regular reviews (including those of Internet celebrity scientists ;-). This is particularly true, when they are conduced according to the strict criteria of the Cochrane Collaboration (something that applies to Wang et al's analysis). If the you want to pick the results of the meta-analysis at hand apart, you will thus have to (a) prove that they deliberately ignored studies although those complied to the inclusion criteria (selection bias) or (b) that important studies that have been included were so biased that the overall result of the meta-analysis (which is mostly math) gets skewed.
The scientists included ony human trials and the deadline on which they stopped looking for new studies was September 3, 2013. In other words: Wang et al. don't bother us with rodent data with questionable relevance (e.g. rodents on 70% fructose diets) and they include almost alll studies in their review that have been published in the last couple of most... well, assuming they were available on MEDLINE, EMBASE, and in the Cochrane databases and complied to the following criteria:
"We included clinical interventions that investigated the chronic effect of exchanging isocaloric or hypercaloric oral fructose for a reference carbohydrate on postprandial triglycerides in humans. Comparisons were considered “isocaloric”if oral fructose in the fructose arm was exchanged for the reference carbohydrate in the control arm in an iso-energetic and iso-glucidic manner and“hypercaloric” if the oral fructose in the fructose arm was provided as a supplement to the background diet providing excess energy (E) relative to the background diet alone in the control arm. Trials with less than 7 days follow-up, which lacked an adequate carbohydrate control, or administered IV-fructose were excluded" (Wang. 2014)
It's quite funny to see how the 1259 initial hits were decimated in the review process with 111 being identified as duplicates, 270 not being having human, but hairier subjects,  54 being only case studies, 2 being letters in disguise, 280 papers being reviews, 233 papers having only a general CHO arm, 71 studies with intravenous administration, 127 studies with "unsuitable endpoints" (e.g. measuring the effect on exercise performance), 61 having a study duration < 7 days and two simply being irretrievable in full-text form.

Now you may be asking yourselves, why I am bothering you with this!? Right? Well, firstly, I want to give yo an idea of how painful it is to write an objective review of the literature. I realized the same only recently, when I compiled the True or False item on dairy induced reductions in testosterone and its possible carcinogenicity. Secondly mentioning the fact that 1211 articles were excluded in the 1st and 48 articles in the 2nd phase of the review process may help silencing all the fructose haters who read this and consider it the work of diabolical cherry pickers, who have received grants from the devil, i.e. the Coca-Cocal Company and a whole host of other usual and unusual suspects, in the past (read the long list of "competing interests" and don't forget that the study at hand was not funded by any high fructose corn-syrup interest group).
A note on potential bias: As I have pointed out numerous times, already. A "competing interest" is no reason to discard the results of a paper / review altogether. Especially in the case of the latter, you should yet carefully evaluate the scientists interpretation of the reviewed literature, because - consciously or not - those interpretations may well be influenced and the corresponding conclusions biased by a researchers' basic assumptions. Unbiased research is - and I am sorry to say that - an illusion that's never going to manifest in the real world; and that's true irrespective of funding / research grants (Schulz. 1995).
While I do hope t hat I am not implying we were talking about scientific fraud here, you should still keep in mind that information Sievenkemper, who is the "correspond author", i.e. the media guy among the 15 scientists from 12 research institutes in Canada, gave Leslie Shepherd, the author of the the initially mentioned press release (Shepherd. 2013), is not some sort of objectively measured truth (there are philosophers of science who question such a thing does even exist). 
"[F]uctose doesn’t behave any differently than other refined carbohydrates. The increases you see are when fructose provides extra calories." (Sievenkemper in Shepherd. 2013)
The above is his (and his colleagues) professional opinion, of which I seriously doubt that it was consciously influenced by previous research grants or the current financial support from the Canadian Institutes of Health Research and the Calorie Control Council that funded the study at hand.
Effects of hypercaloric diet (+50%) w/ 30% fructose content on triglyceride production and clearance in healthy subjects in the presence and absence of exercise (Egli. 2013)
A minimalist explanation of the results: Based on the way fructose is metabolized (increased triglyceride production, reduced storage in fat cells; cf. Chong. 2007), it is only logical that there is a minimal effect on serum triglycerides. In the absence of a hypercaloric diet, this statistically and physiologically non-significant increase is yet not a threat to your health. Your body will simply use the part of fructose your liver converts to triglycerides as an energy source. Only when the total energy intake is so high that it is no longer necessary / possible to use the trigs as an energy source, the latter will begin to accumulate in the blood and. even worse, in the liver (NAFLD). For you that would mean trouble - unless, of course, you work out and use the superfluous trigs to fuel your workouts (Egli. 2013; figure to the left).
Basically, what the scientists did to form this "professional opinion" was (1) reading the papers several times, (2) weighing them by a set of pre-determined criteria from the Cochrane Handbook for systematic reviews of interventions (Higgins. 2011), (3) filtering out all relevant data, (4) calculating the SMD's (standard mean differences) for pre vs. post intervention triglycerides levels of the average study subject for each individual study, (5) pooling the data in groups (healthy subjects, overweight / obese subjects, diabetics), and finally (6) using the individual weight of the to calculated the SMDs [including 95% confidence intervals] for different subject groups. The results, i.e....
  • Complete results (Wang. 2014)
    otherwise health: 0.30 [-0.00, 0.60]; weight 37.8%
  • overweight / obese: 0.69 [0.20, 1.19] *; weight: 6.8%
  • diabetes: 0.00 [-0.15, 014]; weight: 55.4%
did then 7) serve as the basis for the magic overall SMD of 0.14 and confidence intervals of [-0.02, 0.30] ,which tell you that the 14% increase in triglycerides is statistically not significant (for someone who does not sit around all day, the same can probably be said for the physiological relevance - specifically in view of the fact that we have no reason to believe that this was not a new steady state; or, more straight forward: It's unlikely that the levels kept increasing after a short adaptation phase.
How realistic are these studies, anyway? Currently the dietary fructose intake of the average fructose intake of fructose is estimated to be contribute 10-15% to our dietary energy intake (Vos. 2008). If we do the math on only those two trials with corresponding fructose intakes, i.e. Huttenen et al. (1976; healthy subjects) and Anderson et al.(1989; diabetic subjects), we get a standard mean differences of 0.019 with 95% confidence intervals of [95% CI: -0.32, 0.35]. The 1.9% increase in postprandial triglycerides the researchers detected in these studies is thus physiologically irrelevant  and statistically in- significant.
Bottom line: There is little doubt that the researchers' conclusion that "fructose has adverse effects only insofar as it contributes to excess calories" (Sievenkemper in Shephard. 2013) is supported by ...
  1. the absence of differences between diets that delivered up to 25% of the daily energy from fructose or other carbohydrate sources, respectively, as well as
  2. the fact that only studies that employed hyper-caloric diets had significant negative effects on the postprandial triglyceride levels (SMD: 0.65 [95% CI: 0.30, 1.01])
Nevertheless, Sievenkemper's comment in the previously cited press release lacks the most important piece of information, i.e. the fact that the potential adverse effects of fructose are not restricted to increases in serum triglycerides and that a similar verdict of acquittal from a peer-reviewed, up-to-date meta-analysis of its effects on the development of NAFLD is still (over-)due.

What we do have, are meta-analysis for it's effects on blood pressure (Ha. 2012), weight gain (Sievenpiper. 2012) and glucose metabolism in diabetics (Cozma. 2013) which argue that replacing glucose with an isocaloric amount of fructose does not affect blood pressure or weight gain and can actually "improve long-term glycemic control, as assessed by glycated blood proteins, without affecting insulin in people with diabetes" (Cozma. 2013).
References:
  • Anderson, J. W., Story, L. J., Zettwoch, N. C., Gustafson, N. J., & Jefferson, B. S. (1989). Metabolic effects of fructose supplementation in diabetic individuals. Diabetes Care, 12(5), 337-344.
  • Bes-Rastrollo M, Schulze MB, Ruiz-Canela M, Martinez-Gonzalez MA (2013) Financial Conflicts of Interest and Reporting Bias Regarding the Association between Sugar-Sweetened Beverages and Weight Gain: A Systematic Review of Systematic Reviews. PLoS Med 10(12): e1001578.
  • Chong, M. F., Fielding, B. A., & Frayn, K. N. (2007). Mechanisms for the acute effect of fructose on postprandial lipemia. The American journal of clinical nutrition, 85(6), 1511-1520.
  • Cozma, A. I., Sievenpiper, J. L., de Souza, R. J., Chiavaroli, L., Ha, V., Wang, D. D., ... & Jenkins, D. J. (2012). Effect of Fructose on Glycemic Control in Diabetes A systematic review and meta-analysis of controlled feeding trials. Diabetes care, 35(7), 1611-1620. 
  • Egli, L., Lecoultre, V., Theytaz, F., Campos, V., Hodson, L., Schneiter, P., ... & Tappy, L. (2013). Exercise Prevents Fructose-Induced Hypertriglyceridemia in Healthy Young Subjects. Diabetes.
  • Ha, V., Sievenpiper, J. L., de Souza, R. J., Chiavaroli, L., Wang, D. D., Cozma, A. I., ... & Jenkins, D. J. (2012). Effect of Fructose on Blood Pressure A Systematic Review and Meta-Analysis of Controlled Feeding Trials. Hypertension, 59(4), 787-795.
  • Huttunen, J. K., MÄkinen, K. K., & Scheinin, A. (1976). Turku sugar studies XI: Effects of sucrose, fructose and xylitol diets on glucose, lipid and urate metabolism. Acta Odontologica, 34(6), 345-351.
  • Schulz, K. F., Chalmers, I., Hayes, R. J., & Altman, D. G. (1995). Empirical evidence of bias. JAMA: the journal of the American Medical Association, 273(5), 408-412.
  • Sievenpiper, J. L., de Souza, R. J., Mirrahimi, A., Matthew, E. Y., Carleton, A. J., Beyene, J., ... & Jenkins, D. J. (2012). Effect of Fructose on Body Weight in Controlled Feeding TrialsA Systematic Review and Meta-analysis. Annals of Internal Medicine, 156(4), 291-304.
  • Shepherd, L. (2013)   Researchers say fructose does not impact emerging indicator for cardiovascular disease. St. Michael's | Newsroom | Our News. < http://www.stmichaelshospital.com/media/detail.php?source=hospital_news/2013/20131230_hn > retrieved on Jan. 01 2014.
  • Vos, M. B., Kimmons, J. E., Gillespie, C., Welsh, J., & Blanck, H. M. (2008). Dietary fructose consumption among US children and adults: the Third National Health and Nutrition Examination Survey. The Medscape Journal of Medicine, 10(7), 160.
  • David Wang, D., Sievenpiper, J. L., de Souza, R. J., Cozma, A. I., Chiavaroli, L., Ha, V., ... & Jenkins, D. J. (2014). Effect of fructose on postprandial triglycerides: A systematic review and meta-analysis of controlled feeding trials. Atherosclerosis, 232(1), 125-133.