Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies

BMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.e7492 (Published 15 January 2013)
Cite this as: BMJ 2013;346:e7492

Recent rapid responses

Rapid responses are electronic letters to the editor. They enable our users to debate issues raised in articles published on bmj.com. Although a selection of rapid responses will be included as edited readers' letters in the weekly print issue of the BMJ, their first appearance online means that they are published articles. If you need the url (web address) of an individual response, perhaps for citation purposes, simply click on the response headline and copy the url from the browser window.

Displaying 1-10 out of 10 published

The conclusion that weight gain with increased (simple) sugar consumption is calorie-driven seems a little hasty, as your comparative energy source is polysaccharides or complex sugars, which are ultimately broken down into the very same monosaccharides they replace. In order to substantiate your conclusion, isoenergetic exchange should be between sugar and another macronutrient, protein or fat.

Competing interests: None declared

Dora L Beluska, Family doctor

Lachine Hospital, The Lachine Hospital 650 16th Avenue Lachine, Quebec H8S 3N5

Click to like:

Recent articles highlight unresolved debate over the role of sugar in the etiology of obesity.1,2 The debate seems to have lost sight of the fact that, even though energy intake in excess of energy expenditure causes obesity, ultimately, it is insufficient fat oxidation that results in obesity.

It is a basic tenet of biochemistry and well-established physiologic phenomenon, that when fat and carbohydrate are consumed together, the carbohydrate is preferentially oxidized. This phenomenon occurs even during exercise, regardless of the type of carbohydrate, even if only a little carbohydrate is consumed. Fat and carbohydrate consumed together in energy dense meals, food, or beverages result in excess energy intake at the same time as suppressed fat oxidation, favoring a positive fat balance.3-5

The articles by Te Morenga et al and Watts do not mention direct suppression of fat oxidation as a potential reason why sugar might promote obesity.

An incomplete view of the underlying etiology is problematic, because interventions and policy that aim to stop obesity by limiting fat only or carbohydrate only, or worse, a particular type of fat or carbohydrate only, will likely not harness enough of the biochemical and physiologic mechanism to be effective. They are doomed to fail if the background meal pattern and lifestyle continue to prioritize carbohydrate oxidation over fat oxidation.

1. Te Morenga L, Mallard S, Mann J. Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies. BMJ 2013; 346:e7492.
2. Watts G. Sugar and the heart: old ideas revisited. BMJ 2013;346:e7800 (15 January).
3. Hellerstein MK. No common energy currency: de novo lipogenesis as the road less traveled. Am J Clin Nutr. 2001;74:707–8.
4. McDevitt RM, Bott SJ, Harding M, Coward WA, Bluck LJ, Prentice AM. De novo lipogenesis during controlled overfeeding with sucrose or glucose in lean and obese women. Am J Clin Nutr. 2001;74:737–46.
5. Stookey JD, Hamer J, Espinoza G, Higa A, Ng V, Tinajero-Deck L, Havel PJ, King JC. Orange juice limits postprandial fat oxidation after breakfast in normal weight adolescents and adults. Adv. Nutr. 3: 629S–635S, 2012.

Competing interests: None declared

Jodi JD Stookey, Nutrition Epidemiologist

CHORI, 5700 Martin Luther King Jr Way, Oakland CA 94609

Click to like:

Dear BMJ Editor:

We are writing to bring to your attention what we believe to be a calculation error contained in the article by Te Morenga et al., entitled “Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies” [1].

In Figure 6 relating to the effect of reducing free sugars on measures of body fatness in children, it appears that an incorrect formula was used for the calculation of an effect size for the James, et al. (2004) study [2]. In James et al. (2004) [2], the data were analyzed using raw BMI and reported with cluster (not person) as the unit of observation (that is, the numbers reported therein are based on the within-cluster means for 29 clusters). We believe the 0.39 reported in [1] was obtained by inadvertently using the estimated among-cluster standard deviation in the denominator of the standardized mean difference. We believe that the correct approach, which can be found in Hedges [3], entails using the estimated within-treatment, among-subject standard deviation as the denominator. Use of this correct standard deviation would yield a standardized effect size of roughly 0.086 instead of 0.39 as reported in [1].

Sincerely,

David Allison, Ph.D.-1
Karen Keating, Ph.D.-1
Kathryn Kaiser, Ph.D.-1
James Shikany, Dr.PH, PA-C-2

University of Alabama at Birmingham
1-School of Public Health, 2-School Of Medicine
Birmingham, AL, USA
Contact: dallison@uab.edu

References:

1) Te Morenga L, Mallard S, Mann J. Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies. BMJ 2013; 346:e7492.

2) James J, Thomas P, Cavan D, Kerr D. Preventing childhood obesity by reducing consumption of carbonated drinks: cluster randomised controlled trial. BMJ 2004; 328(7450):1237.

3) Hedges LV. Effect sizes in cluster-randomized designs. J Educ Behav Stat 2007; 32(4):341.

Competing interests: Dr. Allison has received financial and other benefits from the following entities: the Frontiers Foundation; The Federal Trade Commission; Kraft Foods; University of Wisconsin; University of Arizona; Paul, Weiss, Wharton & Garrison LLP; Sage Publications, and additional government, non-profit and for-profit organizations with interests in obesity, nutrition, and health. The other authors have no competing interests to declare.

David B. Allison, Associate Dean, Distinguished Professor

Karen Keating, Kathryn Kaiser, James Shikany

University of Alabama at Birmingham, 1720 2nd Ave South, RPHB 140J, Birmingham, AL, USA 35294

Click to like:

The meta-analysis was primarily intended to determine what effect on body weight might be expected should free-living individuals increase their intake of sugars or foods and drinks containing sugars, or be advised to decrease their intake. It was therefore appropriate to exclude studies in which changes in other dietary or exercise practices were recommended, and to include studies where total energy intake was not prescribed. Inevitably in such studies some change in protein, fat and energy occurred, but rather than confound our results, this enabled us to answer the practical question we had posed.

Saris et al. (2000) compared three diets. Dr Cottrell appears to have assumed that we compared the high sugars diet with the “control” diet. This was not the case. The “control” group followed a relatively high fat diet. Given the issue under examination, we appropriately compared the two other diets with similar proportions of fat and total carbohydrate; one being relatively higher and the other relatively lower in simple carbohydrate (sugars). Comparison with the group consuming a higher fat intake was not germane to the meta-analysis.

The heterogeneity, potential publication bias, and absence of a dose response effect are all discussed in our paper. However, the remarkably similar weight change when intake of sugars is increased or decreased over relatively short periods and the suggestion that the effect is more marked when dietary change is implemented over a longer period suggests that advice to limit sugars intake might usefully be included amongst the raft of measures needed to battle the global epidemic of obesity.

Reference

Saris WHM, Astrup A, Prentice AM, Zunft FJF, Formiguera X, Verboeket-van de Venne WPHG, Raben A, Poppitt SD, Seppelt B, Johnston S, Vasilaris TH, Keogh GF (2000) Randomized controlled trial of changes in dietary carbohydrate/fat ratio and simple vs complex carbohydrates on body weight and blood lipids: the CARMEN study. International Journal of Obesity 24:1310-1318.

Competing interests: None declared

Lisa Te Morenga, Research Fellow

Jim Mann, Simonette Mallard

University of Otago, PO Box 56 Dunedin, 9054 New Zealand

Click to like:

We read with interest the high-quality systematic review by Te Morengo et al. [1]. Although the conclusion that sugars in isocaloric exchange with other sources of carbohydrate do not affect body weight is well supported by the available evidence and agrees well with the effect of sugars on other related cardiometabolic endpoints, the conclusion that sugars are a determinant of body weight from ad libitum trials requires some attention.

A large database of feeding trials have studied the effect of sugars in isocaloric exchange for other sources of carbohydrate. We have conducted a series of systematic reviews and meta-analyses of these trials in relation to the effect of fructose (the presumed culprit) on cardiometabolic risk factors (ClinicalTrials.gov identifier: NCT01363791). In agreement with the present systematic review [1], pooled analyses have shown that fructose in isocaloric exchange for other carbohydrates does not increase body weight [2]. The same is also true for fasting lipids [3] and uric acid [4], as well as postprandial lipids and non-alcoholic fatty liver disease (NAFLD) [unpublished data] with benefits seen for glycemic control [5] and blood pressure [6]. In these syntheses, fructose was no worse than other sources of carbohydrate (even under positive energy balance, at large doses well above the mean level of fructose intake [7],or in fluid form) as long as the comparisons remain matched for energy. As more trials have become available, earlier signals for harm in relation to fasting (>60 g/day [3] and >100 g/day [8]) and postprandial (>50 g/day [8]) triglycerides at high-dose thresholds have not been replicated [unpublished data].

The authors interpreted the ad libitum trials as providing evidence for harm [1]. In the absence of an effect under conditions of isocaloric exchange of sugars with other sources of carbohydrate, it was concluded that changes in body fatness are mediated through the ability of sugars in ad libitum trials (defined as, “no strict control of food intake”) to increase energy intake rather than any metabolic effects particular to sugars. The problem is whether the design of these trials allows this inference to be drawn. A true ad libitum design should allow for the free selection of study foods (sugary foods and beverages) not just the background diets. The trials in question required that the participants consume all sugary study foods and beverages providing excess energy or eliminate these foods while not strictly controlling the food intake of the background diets. The comparator was the background diets alone [1]. In the absence of a macronutrient comparator providing or displacing the same amount of excess energy, this design is necessarily an imbalanced, hypercaloric (more calories in the sugar arm) design which makes it difficult to conclude that sugary foods and beverages more than any other caloric foods lead to greater overall caloric intake and weight gain. This is the interpretation we adopted with our systematic reviews and meta-analyses of the effect of fructose on cardiometabolic risk factors. As with the present review [1], we found a consistent signal for harm in hypercaloric trials in which diets were supplemented with fructose providing excess energy (+18-97% energy) at extreme doses (+104-250 g/day) well above the 95th-percentile for intake in the population [7] compared with the same diets alone (without the excess energy). Under these conditions, pooled analyses showed clinically important increases in body weight and uric acid [2,3], as well as in fasting triglycerides, postprandial triglycerides, markers of non-alcoholic fatty liver disease (NAFLD), glucose, and insulin [unpublished data]. When the excess energy from fructose in these trials, however, was matched with excess energy from glucose, the signal for harm disappeared [2-6]. That is, fructose did not behave differently than glucose as a source of excess energy. The implication is that the observed adverse effects of sugars and fructose on body weight and cardiometabolic risk may be more attributable to the provision of excess energy rather than the sugar itself.

Before one can conclude that fructose-containing sugars uniquely lead to an increase in overall energy intake, weight gain and greater cardiometabolic risk, evidence of a meaningful effect of sugars on energy intake and weight gain beyond that of other commonly consumed forms of energy (such as refined grains, salty snacks, or processed meats) must first be established. As the decrease in energy from one food will tend to be compensated by the increases in energy from another food to maintain energy balance under free living, ad libitum conditions, it is important to compare fructose-containing sugars with those foods likely to replace these sugars in the diet. True ad libitum trials which compare fructose-containing sugars with other sources of energy under free living, “real world” conditions are urgently needed.

References
1. Te Morenga L, Mallard S, Mann J. Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies. BMJ 2013;346:e7492
2. Sievenpiper JL, de Souza RJ, Mirrahimi A, Yu ME, Carleton AJ, Chiavaroli L, DiBuono M, Jenkins AL, Leiter LA, Wolever TWS, Beyene J, Kendall CWC, Jenkins DJA. Effect of fructose feeding on body weight: systematic review and meta-analyses of controlled feeding trials. Ann Intern Med 2012;156, 291-304
3. Sievenpiper JL, Carleton AJ, Chatha S, Jiang HY, de Souza RJ, Beyene J, Kendall CW, Jenkins DJ. Heterogeneous effects of fructose on blood lipids in individuals with type 2 diabetes: systematic review and meta-analysis of experimental trials in humans. Diabetes Care. 2009;32:1930-7
4. Wang DD, Sievenpiper JL, de Souza RJ, Chiavaroli L, Ha V, Cozma AI, Mirrahimi A, Yu ME, Carleton AJ, DiBuono M, Jenkins AL, Leiter LA, Wolever TWS, Beyene J, Kendall CWC, Jenkins DJA. The effects of fructose intake on serum uric acid vary among controlled dietary trials. J Nutr. 2012;142:916-23.
5. Cozma AI, Sievenpiper JL, de Souza RJ, Chiavaroli L, Ha V, Wang DD, Mirrahimi A, Yu ME, Carleton AJ, DiBuono M, Jenkins AL, Leiter LA, Wolever TWS, Beyene J, Kendall CWC, Jenkins DJA. Effect of Fructose on Glycemic control in Diabetes: A Systematic Review and Meta-Analysis of Controlled Feeding Trials. Diabetes Care. 2012;35:1611-20.
6. Ha V, Sievenpiper JL, de Souza RJ, Chiavaroli L, Wang DD, Cozma AI, Mirrahimi A, Yu ME, Carleton AJ, DiBuono M, Jenkins AL, Leiter LA, Wolever TWS, Beyene J, Kendall CWC, Jenkins DJA. Effect of Fructose on Blood Pressure: A Systematic Review and Meta-Analysis of Controlled Feeding Trials. Hypertension. 2012;59:787-95.
7. Marriott BP, Cole N, Lee E. National estimates of dietary fructose intake increased from 1977 to 2004 in the united states. J Nutr. 2009;139:1228S-1235S.
8. Livesey G, Taylor R. Fructose consumption and consequences for glycation, plasma triacylglycerol, and body weight: meta-analyses and meta-regression models of intervention studies. Am J Clin Nutr. 2008;88:1419-37.

Competing interests: I have received research support from the Canadian Institutes of health Research (CIHR), Calorie Control Council, The Coca-Cola Company (investigator initiated, unrestricted grant), Pulse Canada, and International Nut Council. I have also received travel funding, speaker fees, and/or honoraria from The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health (NIH), Canadian Diabetes Association (CDA), Calorie Control Council, Diabetes and Nutrition Study Group (DNSG) of the European Association for the Study of Diabetes (EASD), International Life Sciences Institute (ILSI) North America, ILSI Brazil, Abbott Laboratories, Pulse Canada, and The Coca-Cola Company. My wife is an employee of Unilever Canada.

John L Sievenpiper, Knowledge Synthesis Lead

Toronto 3D Knowledge Synthesis and Clinical Trials Unit, Clinical Nutrition and Risk Factor Modification Centre, St. Michael's Hospital, #6137-61 Queen Street East, Toronto, ON, M5C 2T2, CANADA

Click to like:

It is always very difficult to study individual components of the human diet. The quest to uncover the ultimate cause of obesity and/or cardiovascular disease seems to hold much interest in the scientific world. Ultimately the public become confused by so many conflicting theories.

Sugar, unlike fat is not an essential nutrient. Sugar is only added to food in the same way as salt and pepper to make it more palatable. Sugar content is particularly high in food aimed for consumption by children. Do children only consume 10% of their diet as sugar, I very much doubt it.

Glucose is of course ultimately the human bodies’ fuel; excess glucose unused by activity is stored as fat. All carbohydrate is converted to glucose. The problem with high sugar diets is they cause insulin spikes followed by a rapid low blood glucose level which often leads to a need for another sugar rush. This undoubtedly leads to problems such as insulin resistance and abdominal obesity. People lead erratic lives with lack of consistency such as regular meal times and expect good health. Consumption of fats and proteins slow digestion and so aid satiety, because they don’t contain glucose/sugar, they don't contribute to erratic blood sugar levels. A balanced diet is paramount to good health. It is such a travesty that the essential component of fat in the diet has been given such a bad reputation.

The study by Morenga et al (1) reviews ‘free living people involving ad libitum diets’ – this is the crux of the problem; diet and lifestyle need to be major considerations, we can’t just do whatever we like. A more holistic approach on how to achieve optimal individual health is urgently required.

(1) Morenga Lisa Te, Mallard Simonette, Mann Jim. Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies BMJ 2013;346:e7492

Competing interests: None declared

Jane E Collis, Independent Researcher

No affiliation, Kenilworth Warks UK

Click to like:

If about 10% of our energy intake is from added sugars - and it is often more - then advising patients to reduce this brings benefits which I have seen in many patients in primary care. To give up sugar in hot drinks and give up soft drinks should be the opening gambit of any sensible dietary advice.

While the parallel between advising obese patients about sugar and advising patients with COPD about smoking is clear, the big difference is that sugar can easily be reduced by everyone as its overuse is almost universal. Replacing empty sugar calories with nutrient rich foods can benefit the thin.

We need government action on this matter as soon as possible, but with the food lobby as powerful now as the tobacco lobby was not so long ago, I won't be betting my pension on it!

Competing interests: None declared

F Colin Francis Bannon, GP

The Mannamead Surgery, Plymouth

Click to like:

The clear evidence that sugar ( in all its numerous forms ) is connected to obesity and diabetes is not new but what is constantly surprising is that few experts seem to appreciate the similar metabolic affect of starch. The body really does not distinguish between white bread or potatoes and sugar. Even Walter Willet has spoken out against starch at various conferences in the US.

Reducing sugar alone will not solve the obesity problem and all the associated health problems such as diabetes.

Competing interests: I am the founder of a company that makes natural low sugar and starch foods

Hannah Sutter, Food Manufacturer

Natural Ketosis, 45-47 Tower Street EH67BN

Click to like:

The review by Te Morenga et al., (2013) sheds some light on the confused literature regarding dietary sugars and body weight. It is refreshing that the authors acknowledge that the relatively small change (~0.8 kg) in body weight attributed to alterations in sugars intake in the intervention trials is probably due to changes in energy intake and not some other characteristic unique to sugars. Their conclusion that sugars intake is a ‘determinant’ of body weight could thus be applied equally to any caloric nutrient or food.

The disparity of the studies does not allow conclusions to be drawn specifically for consumption of ‘free sugars’, which appeared to be the question posed to the authors. Since the review included the results of studies which changed consumption of sugars-containing foods, the results are confounded by changes to other caloric nutrients. The data did not allow derivation of a dose response and provide no direct support for the controversial suggestion that average population “free sugars” intake should be limited to 10% of food energy.

Some important points need to be raised with regard to what appears to be evidence of selection bias in the studies and data included in the review. Firstly, the only data used from the CARMEN Study (Saris et al., 2000) is incorrectly included in Table 2/Figure 3 (effect of reducing sugars on body weight in adults). The main result of this long term intervention should have been included in Table 3/Figure 4 (effect of increasing sugars on body weight in adults). This trial showed that increasing sugars to 30% energy for six months gave rise to a fall in body weight of 1.7 kg compared to control.

Second, the data included in Table 3/Figure 4 from the Poppitt et al. study (that was an adjunct of the CARMEN study, but involved subjects with metabolic syndrome) is the comparison between the high sugars group and the high starch group. The correct comparison should have been between the high sugars group and the controls. This shows no significant change in body weight after six months on a diet again containing 30% energy as sugars, when compared with the subject’s usual intake.

Third, the correct citation for the study by Marckmann et al., (2000) is Raben et al., (1997) [1], which reported the body weight results of this study. The present review arbitrarily cites part of the results of this study (the comparison of the high sugars group with one of the low sugars groups) while ignoring the non-significant changes in body weight seen when the high sugars group was compared to the other low sugar group.

These errors would be expected to affect the reported effect sizes and confidence intervals.

The heterogeneity of the interventions, and the admitted evidence of publication bias in the studies available for the meta-analysis, diminishes any confidence in the authors’ conclusions. Indeed, the results of studies where dietary sugars were isoenergetically exchanged for other carbohydrates or macronutrients show no effect on body weight (Fig 5). This finding, together with absence of evidence for a dose response, concur with reports by the Institute of Medicine [2], European Food Safety Authority [3], National Health and Medical Research Council [4], and the earlier joint FAO/WHO Consultation on Carbohydrates in Human Nutrition [5] that were unable to identify any level of added or total sugars intake likely to increase energy intake or obesity risk.

As the authors acknowledge, the effectiveness of the application of the results of this present analysis to the general public remain to be demonstrated.

1. Raben A., Macdonald I., Astrup A. (1997) Replacement of dietary fat by sucrose or starch: effects on 14 d ad libitum energy intake, energy expenditure and body weight in formerly obese and never-obese subjects. International journal of obesity and related metabolic disorders. Journal of the International Association for the Study of Obesity, 21, 846-59
2. Food and Nutrition Board, Institute of Medicine of the National Academies (2005) Dietary Reference Intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, amino acids (macronutrients). The National Academies Press. Washington, DC.
3. European Food Standards Authority (2010) Scientific opinion on dietary reference values for carbohydrates and dietary fibre. EFSA Panel on Dietetic Products, Nutrition and Allergies. EFSA Journal 2010; 8(3):1462.
4. National Health and Medical Research Council (2003) Dietary Guidelines for Australian Adults. The Commonwealth of Australia, Canberra
5. FAO/WHO (1997) Carbohydrates in human nutrition (FAO Food and Nutrition Paper - 66). FAO, Rome

Competing interests: The authors work for the World Sugar Research Organisation - a non-profit organisation supported by the sugar industry.

Richard C Cottrell, Director-General

Anna Wittekind

World Sugar Research Organisation, 70 Collingwood House, Dolphin Square, London SW1V 3LX

Click to like:

17 January 2013

Honey and sugar are addictive sweets that trick and trap us by creating the fleeting euphoria of sweetness and peace, but the sustained sickness of bitterness and anger. The euphoria of sweetness and peace, and the sickness of bitterness and anger, are polar opposites that reinforce each other: the euphoria blinds us to the sickness, and the sickness makes us crave the euphoria. Perversely but predictably, honey and sugar create, aggravate, and perpetuate the very sickness of bitterness and anger that they seem to cure, thus placing many favorite sweets in a bad light.

Competing interests: None declared

Hugh Mann, Physician

Retired, Eagle Rock, MO, USA

Click to like:

THIS WEEK'S POLL