sign in sign up
<<
Home , Added sugar, Clinical nutrition

Diabetes Care Volume 37, April 2014 957 YE2DAEE PREVENTION DIABETES 2 TYPE

1 2,3 Dietary and Body Weight: Richard Kahn and John L. Sievenpiper Have We Reached a Crisis in the Epidemic of Obesity and Diabetes? We Have, but the Pox on Sugar Is Overwrought and Overworked Diabetes Care 2014;37:957–962 | DOI: 10.2337/dc13-2506

In the preceding point narrative, Drs. Bray and Popkin provide their opinion and review data that suggest to them that we need to reconsider the consumption of dietary sugar based on the growing concern of obesity and . In the counterpoint narrative below, we argue that there is no clear or convincing evidence that any dietary or added sugar has a unique or detrimental impact relative to any other source of calories on the development of obesity or diabetes. Sugar is purely a highly palatable source of energy; because it has no other property that appears to contribute to our nutritional well-being, it is not an essential for most of us. For those who wish to reduce energy consumption, ingesting less sugar is a good place to start. However, doing so does not automatically portend any clinical benefit.

In this counterpoint discussion, we use the phrase “dietary sugar” or “added sugar” to mean or high- corn (HFCS). Almost all dietary or added sugar used as an ingredient in either solid (e.g., , snacks) or liquid (e.g., sugar-sweetened beverages [SSB]) is in the form of these two . Although we will discuss evidence from feeding studies in which fructose itself was used as the sole added sweetener, it should be noted that fructose rarely occurs alone in foods commonly consumed by humans. Also of importance is the fact that sucrose and HFCS are both composed of and fructose. Whereas the ratio of glucose to fructose is equal in sucrose, in HFCS the ratio is usually 55% fructose, 42% 1 glucose, and 3% glucose polymers; other forms of HFCS have a lower proportion of Department of , University of North Carolina School of Medicine, Chapel Hill, NC fructose. In addition, the glucose and fructose in HFCS are free in solution; in sucrose 2Toronto 3D Knowledge Synthesis and Clinical they are initially bound together. But when sucrose is used in processed or prepared Trials Unit, Clinical and Risk Factor foods/beverages an appreciable amount is broken down to free fructose and glu- Modification Centre, St. Michael’sHospital, cose prior to consumption. Finally, whereas glucose and fructose are metabolized Toronto, Canada 3Department of and Molecular Medi- differently, the belief that sucrose is metabolized differently than HFCS is a myth. No cine, Faculty of Health Sciences, McMaster Uni- study has shown any difference between the two when each is given isocalorically, versity, Hamilton, Canada – nor is there any difference in or caloric value (1 3). Corresponding author: Richard Kahn, rak6200@ Much of the condemnation of sugar in the last few years owes its origin to an gmail.com. article by Bray et al. (4) showing an ecological relationship between sugar availability © 2014 by the American Diabetes Association. (a crude measure of intake) and obesity, which has now been expanded to explain a See http://creativecommons.org/licenses/by- myriad of metabolic abnormalities (5–7). Table 1 shows that the rise in the preva- nc-nd/3.0/ for details. lence of overweight/obesity in the early 1980s does indeed appear to be related to See accompanying articles, pp. 906, an increase in the availability of added . However, starting around 2000, sugar 909, 912, 922, 934, 943, and 950. 958 Point-Counterpoint Diabetes Care Volume 37, April 2014

Table 1—Prevalence of overweight/obesity and its relationship to estimated dietary property of sugar, or simply because total sugar consumption energy consumption in the intervention Prevalence of Estimated intake Estimated intake Estimated was less than in control subjects. overweight/ of added sugars of caloric sweeteners intake of HFCS Finally, all four research groups (9–12) Years obesity (%)* (% daily energy) (kcal/day)**** (kcal/day)**** meta-analyzedtrialsinwhichanin- 1971–1974 46.5 NA 411 7 creased amount of calories from sugar 1976–1980 46.5 13.1** 409 36 were given to adults (there were no 1988–1994 55.9 13.5** 448 170 studies in children) as a supplement to 1999–2000 64.5 18.1*** 502 211 their normal . The summary point estimates in all four reports showed a 2007–2008 68.0 14.6*** 457 189 modest but significant weight gain even 2009–2010 68.7 NA 440 169 though in most of the individual studies *Data obtained from references 64–67. **Data obtained from reference 68. ***Data obtained the CI crossed unity. Again, whether the from reference 69. ****Data obtained from reference 70. change in weight was due to the increase in energy consumed (as would be ex- pected in the absence of complete com- consumption appears to have declined affect body weight in adults. In both pensation) or to some unique property considerably, but the prevalence of obe- analyses the forest plot summary esti- of sugar is unknown. sity (and diabetes, data not shown) has matesshowednosignificant effect of Overall, therefore, when sugar was continued to rise. Such ecological find- sugar or pure fructose on body weight replaced in an isocaloric exchange, there ings are certainly intriguing, but because (relative risk [RR] 0.26 [95% CI 20.26 to was no change in body weight. When obesity appears related to many changes 0.83] and RR 20.18 [95% CI 20.47 to subjects were randomized to receive in our environment (8), including a rise in 0.23], respectively). No isocaloric trials fewer calories from sugar and thus con- the consumption of bottled (9), have been performed in children. sumed less total energy, the studies they are only hypothesis-generating. Another approach in RCTs has been to generally showed no significant weight The preferred methodology to deter- examine the effect on weight when cal- loss or were sufficiently confounded to mine whether sugar itself causes weight ories from sugar are reduced relative to preclude reaching any conclusions. Con- gain would be to conduct a randomized consumption in the control group. Te versely, adults given added energy in controlled trial (RCT), in which the con- Morenga et al. (13) meta-analyzed five the form of sugar gained weight. As sumption of sugar is the only variable such trials in children and the summary weight gain or loss only occurred when between otherwise identical groups. point estimate was not significant. Kaiser sugar (energy) was added to, or reduced RCTs are, of course, the strongest form et al. (9) also found no significant change from, the usual background diet, and of evidence (10,11) because they elimi- in weight with a reduction in calories there was no change in weight from an nate the possibility of confounding from from sugar when eight trials in children isocaloric exchange, it seems likely that prerandomization factors and can de- and adults were meta-analyzed to- any effect of sugar on weight is because termine the effect of sugar independent gether. However, a meta-analysis by of the energy it supplies and not be- of other dietary components. Malik et al. (14) identified two of five cause it has any unique property. trials that showed a significant loss of Unfortunately, virtually all the indi- RCTS ON THE EFFECT OF SUGAR ON weight with a reduction in calories from vidual trials meta-analyzed in the four BODY WEIGHT sugar, and the summary point estimate reports discussed above, recruited few Over the last decade, numerous RCTs on was significant when a fixed-effects subjects (,100) and the intervention the effects of sugar consumption have model was used but not with a random- was of short duration (,1 year). Publi- been performed. Very recently, four in- effects model. Te Morenga et al. (13) cation bias was also noted in the trials dependent groups have performed sys- also meta-analyzed trials conducted in conducted in adults (11,12). The only tematic reviews and meta-analyses of adults and found that the summary point exceptions to these design issues were these published trials (9,12–14). Each estimate significantly favored a reduc- two recent trials (15,16), where a large of these studies used different inclusion tion in body /weight resulting from a number of children (224 and 641, re- and exclusion rules but in all four the out- reductionincaloriesfromsugarcon- spectively) were randomized to reduced come of interest was weight change. Al- sumption. However, when three of five reduction in calories from sugar con- though each of the four meta-analyzed studies were removed from the anal- sumption relative to control subjects many of the same trials, they often ysis because they had a high risk of and the interventions were carried out grouped them differently and each used bias, the summary point estimate was for 12 and 24 months, respectively. Al- different exclusion/inclusion rules no longer significant. All three meta- though both trials were included in two and reported different summary point analyses (9,13,14) found major interstudy of the above meta-analyses (9,14), their estimates. heterogeneity (I2 .50%). individual results may be instructive. Te Morenga et al. (13) and Sievenpiper Of note, in all the individual trials in Both trials (15,16) reported significant et al. (12) examined whether an isoca- which subjects consumed less calories weight reduction when the consump- loric exchange of added sugar or puri- from sugar they also consumed less to- tion of calories from SSBs was reduced. fied fructose with other macronutrients tal energy, and therefore it is unclear if However, in one study (15) weight loss (mostly other ) would any weight loss was due to some unique was quite modest after 12 months care.diabetesjournals.org Kahn and Sievenpiper 959

(mean difference between groups (14) showed that when adjusted for total from prospective cohort studies ap- 20.13 [95% CI 20.20 to 20.06]). In energy consumed the once positive rela- pears very problematic. the other (16), the prespecified primary tionship was no longer significant. body weight end point was not signifi- Supporting our claim that there is EFFECTS OF SUGAR ON APPETITE cant, but was so (P 5 0.045) at an in- nothing special about calories from AND SATIETY terim 1-year analysis. In a subgroup sugar, many other sources of highly pal- Some investigators argue that an ad- analysis, the positive effect for the pri- atable calories can also increase body verse effect of sugar, particularly when mary end point occurred only in Hispanic weight. For example, in a pooled analysis consumed as SSBs, is that it stimulates children. Thus, these larger and longer of three of the well-known Harvard co- appetite or reduces satiety (4,6,34). duration trials show no definitive adverse horts (which are often cited [5–7,17] as Many investigators have pursued this effect of sugar on body weight, despite showing that sugar causes obesity and hypothesis and the results have been the fact that the intervention groups con- diabetes) an increase in one serving of conflicting. Recently, Almiron-Roig et al. sumed less total energy throughout the French fries (13.35 lbs), potato chips (35) performed a systematic review follow-up period, which would be ex- (11.69 lbs), unprocessed meat (10.93 and meta-analysis of the studies on pected to favor weight loss. lbs), or boiled, baked or mashed pota- this topic. The question they addressed toes (157 lbs) resulted in greater or is whether energy given before a meal PROSPECTIVE STUDIES ON THE similar weight gain as did sugary bever- (i.e., preload) will affect the energy EFFECT OF SUGAR ON BODY ages (11.0 lbs) for every 4 years of follow- consumed at a meal. Their analyses WEIGHT up, when intake was not adjusted for total showed that overeating was much As RCTs do not indicate that sugar con- energy consumption (18). greater with liquid preloads than with sumption itself causes weight gain, why Of note, there are many methodolog- solid or semisolid preloads. Whereas do some investigators believe otherwise? ical problems with the prospective co- this finding might support the claim One possibility is that they focus instead hort studies that are related to sugar that SSBs increase energy consumption, on the results of prospective cohort stud- consumption. First, they obviously suf- the authors also found that the effect of ies. In the latter, subjects are asked to fer from the inability to control for all liquids did not correlate significantly complete a semiquantitative food fre- the variables that could lead to residual with their energy content. That is, it quency questionnaire to ascertain the confounding; indeed, the vast majority was the liquid nature of the preload, consumption of specific foods. Partici- do not adjust for caloric intake. Second, rather than the energy within it, that pants are then followed, often for years, they did not publicly prespecify how ex- influenced subsequent food consump- and many outcomes are recorded. posure would be defined (e.g., quartile- tion. Therefore, the hypothesis that Te Morenga et al. (13) and Malik et al. or tertile-defined categories, highest vs. sugar per se leads to excess food con- (14) also conducted a systematic review lowest), the number of analyses that sumption is not supported by the totality and meta-analysis of such cohort stud- would be performed, or the statistical of the evidence. ies. Their results indicated that a major- tests to be used, and the results were The Almiron-Roig et al. (35) review ity of the studies found a significantly not adjusted for repeated tests of signif- focused only on compensatory energy positive association between sugar in- icance. Any of these problems could intake after a preload. All the studies take and various measures of body have easily led to spurious results. they examined did not report whether weight. Te Morenga et al. (13), however, Finally, the essence of prospective co- overeating translated into persistent noted that because a wide variety of hort studies in nutrition is their reliance weight gain. There appears to be, how- measures of adiposity were used (e.g., on the ability of subjects to recall accu- ever, only two studies that examined BMI, skinfold thickness, incident over- rately exactly what they ingested. In the the effect of liquid versus solid energy weight or obesity, weight, percent context of meals consisting of a variety on weight, and both showed no signifi- body fat), it was difficult to draw a de- of foods assembled in a myriad of ways cant effect of food form on weight finitive conclusion because in many and that often vary over time, along change (36,37). studies some measures were significant with changing tastes, lifestyles, and the All told, therefore, we have no per- whereas others were not. In the Malik constant introduction of new products suasive clinical evidence that sugar in et al. (14) meta-analysis, the studies and packagingdit should not be surpris- beverages enhances energy consump- considered displayed high heterogene- ing that many studies have shown that tion or that liquid energy is weight- ity and publication bias. such questionnaires have substantial promoting any more so than solid Of importance, while all of the pro- biases and inaccuracies (19–25). More- energy. Liquids (e.g., SSBs, milk) ingested spective cohort studies examining the over,evenwhenfocusingonaspecific around or shortly before a meal in exper- relationship between sugar and weight food such as SSBs, such single nutrient imental conditions do seem to stimulate adjusted for various potentially con- analyses may be confounded by dietary overeating more so than energy in other founding variables, almost none ad- pattern; over- or underreporting of in- forms, but whether that translates into justed for energy consumption. Thus, take can also be different depending on long-term weight gain is unknown. in these studies, as in the RCTs, the pos- the demographic characteristics of the itive association between increased population (26–33). For all these rea- WHAT ABOUT FRUCTOSE OR HFCS? sugar consumption and weight could well sons as well as the others mentioned Some investigators have focused on the be due to excess energy intake and not to a above, the claim that sugar itself pro- fructose component of sucrose/HFCS as unique effect of sugar. Indeed, Malik et al. motes weight gain based on the results the “evildoer” of sugar consumption as 960 Point-Counterpoint Diabetes Care Volume 37, April 2014

glucose and fructose are absorbed and another large cohort study published that hypothesis, the Malik et al. (40) ar- metabolized differently (2). The impact earlier by the same research group, ticle discussed above performed a sec- of fructose on body weight was re- but not included in the meta-analysis, ond meta-analysis and concluded that viewed above and the evidence sug- found no association between total SSBs were associated with the develop- gests no difference between fructose sugar intake and diabetes (41). Also, ment of metabolic syndrome. Of note, and any other (12,38). one of the studies included in the however, Malik et al. (40) identified only In addition, as concluded above, if there meta-analysis (42) showed no signifi- three cohort studies that had metabolic is an effect of any sugar on weight, it cant association between intake of SSB syndrome as the outcome of interest. In appears to be because of its contribu- and the development of diabetes when two of the three, the relative risk be- tion to total energy consumed. Addi- the data were adjusted for energy in- tween the extreme quartiles of SSB con- tional evidence for that assertion take, but that finding was not men- sumption was not significant, and in the comes from weight-loss trials, where tioned in the meta-analysis. In that remaining study where the relative risk changes in the macronutrient composi- study (42), the consumption of artifi- cited was significant (61), that result tion of the diet (high or low carbohy- cially sweetened beverages was signifi- pertained the consumption of any soft drate, high or low fat) in the setting of cantly associated with diabetes when (regular or diet) and the data were equally hypocaloric diets result in an adjusted for total energy intake, which not adjusted for , body weight, equivalent weight loss (39). is similar to a recent report (43) showing or energy intake. Moreover, a closer ex- In conclusion, there is no evidence that both SSBs and artificially sweet- amination of the results of the latter that fructose or HFCS per se causes obe- ened beverages conveyed equally signif- study (61) indicate that the risk of met- sity or even weight gain. Sugar obviously icant risk. abolic syndrome was the same when contains energy, and there is some evi- Finally, other prospective cohort regular soda consumption was com- dence, albeit conflicting, incomplete, studies have shown a significant nega- pared with diet soda and at any level and inconclusive, that excess energy tive association for total sugars (44) or of consumption. consumption in the form of any sugar sucrose (45) with diabetes and no signif- The hypothesis that fructose-induced may contribute to weight gain. If the icant positive association for total su- hyperuricemia is harmful was recently excess energy in sugar is the culprit, it crose or fructose (41,44,46), and one tested in a double-blind RCT (62). High is reasonable to conclude that any food study showed a positive association for fructose consumption in an overall iso- consumed in excess is just as likely to total fructose and a negative association caloric diet did indeed result in an alter energy balance as would an equal for sucrose (45). All told, therefore, increase in serum uric acid when com- caloric amount of sugar. On the other thereisnopersuasiveevidencefor pared with an isocaloric high-glucose hand, sugar contains no essential micro- a role of sugar in the development of diet. Yet the high-fructose diet had no nutrient and therefore if a reduction in diabetes. significant effect on a wide variety of energy intake is desirable, reducing Other studies have examined surro- hepatic biomarkers, including triacy- sugar consumption is obviously the gatemeasuresofdiabetesrisk,also glyerol. On the other hand, when either place to start. with mixed results. Most (47–53) but sugarwasgiveninanoverallhyper- not all (54) controlled trials showed caloric diet, they both produced similar SUGAR AND DIABETES that fructose or sucrose had no adverse significantchangesinbiomarkersof It is well accepted that weight gain is a effect on fasting plasma glucose, post- nonalcoholic fatty liver disease. These major risk factor for the development of prandial glucose, or insulin levels. Simi- results suggest, once again, that any diabetes. As dietary sugar itself does not lar inconsistent results have been adverse effect of a sugardparticularly appear to have a significant role in reported on measures of insulin resis- fructosedis due to the excess energy weight gain, it is possible that sugar al- tance regardless if the sugar is sucrose it provides and not the molecule itself. ters in some other regard or fructose (51–57). In addition, meta- Even if one believes that sucrose or thereby causing diabetes. From a clinical analyses of controlled trials have fructose adversely affects some aspect perspective, there are no RCTs examin- shown that fructose administration im- of metabolism related to the develop- ing whether sugar consumption in sub- proves glycemia in people with diabe- ment of diabetes, all of the surrogate out- jects with normoglycemia results in tes (58,59). There appears to be no come studies were conducted over days diabetes or even prediabetes. Prospec- study on the effect of sugar on b-cell or at most a few weeks and therefore any tive cohort studies have generated con- function, which is as important in the adverse effect of sugar certainly does not flicting results. For example, Malik et al. development of diabetes as is insulin mean the diabetes would eventually de- (40) performed a meta-analysis of large, resistance. velop. Also, the vast majority of feeding long-term cohort studies and found a Johnson et al. (60) recently reviewed studies showing an adverse effect of significant association between SSBs the literature on the effect of fructose sugar on a metabolic parameter related and incident diabetes. A close examina- consumption on the development of di- to diabetes gave subjects sucrose or fruc- tion of the eight studies included reveals abetes and obesity. Despite some con- tose in a hypercaloric exchange with that four did not find a significant asso- flicting evidence, they hypothesized other sources of energy, or in addition ciation between SSBs and diabetes. that fructose induces hyperuricemia, to a background diet, or (particularly Moreover, five of the eight did not ad- which then results in the development when studying fructose) in amounts usu- just their findings for energy intake of the metabolic syndrome. Although ally exceeding the 95th percentile of or even body weight. Interestingly, there has been no RCT that has tested consumption (3,63). Conversely, when care.diabetesjournals.org Kahn and Sievenpiper 961

8. McAllister EJ, Dhurandhar NV, Keith SW, Table 2—Putative effects of sugar* et al. Ten putative contributors to the obesity Adverse outcome Strength of the evidence epidemic. Crit Rev Food Sci Nutr 2009;49:868– Increases weight in an isocaloric exchange with other 913 macronutrients None 9. Kaiser KA, Shikany JM, Keating KD, Allison DB. Will - Increases weight in a hypercaloric diet** Moderate/inconsistent consumption reduce obesity? Evidence sup- Decreases weight in a hypocaloric diet** Weak/inconsistent porting conjecture is strong, but evidence Increases appetite resulting in weight gain None when testing effect is weak. Obes Rev 2013; Causes diabetes None 14:620–633 Provides unnecessary energy Strong 10. Atkins D, Best D, Briss PA, et al.; GRADE Working Group. Grading quality of evidence *Sugar defined as sucrose, glucose, fructose, or HFCS. **A hypercaloric diet is defined as a diet in and strength of recommendations. BMJ 2004; which energy intake exceeds energy expenditure, and a hypocaloric diet is defined as a diet in 328:1490 which energy intake is less than energy expenditure, thereby favoring weight gain and weight 11. Guyatt GH, Oxman AD, Kunz R, et al.; loss, respectively. GRADE Working Group. What is “quality of ev- idence” and why is it important to clinicians? BMJ 2008;336:995–998 12. Sievenpiper JL, de Souza RJ, Mirrahimi A, et al. Effect of fructose on body weight in con- sucrose or fructose was given in an overall list because they provide no essential trolled feeding trials: a systematic review and isocaloric diet or at the 50th percentile nutrients. meta-analysis. Ann Intern Med 2012;156:291– of consumption adverse effects have 304 rarely been reported (3,12,63). There- 13. Te Morenga L, Mallard S, Mann J. Dietary fore, whether sugar consumption has Funding. J.L.S. received research grants/support sugars and body weight: systematic review and from the Canadian Institutes of Health Research, even an indirect effect on the develop- meta-analyses of randomised controlled trials Calorie Control Council, The Coca-Cola Company and cohort studies. BMJ 2013;346:e7492 ment of diabetes is quite unclear. (investigator-initiated, unrestricted grant), Pulse 14. Malik VS, Pan A, Willett WC, Hu FB. Sugar- Canada, and The International Tree Nut Council sweetened beverages and weight gain in children Nutrition Research & Education Foundation. CONCLUSIONS and adults: a systematic review and meta-analysis. Duality of Interest. J.L.S. has received speak- Am J Clin Nutr 2013;98:1084–1102 Table 2 summarizes what we believe is er’s fees and honoraria from the American So- 15. de Ruyter JC, Olthof MR, Seidell JC, Katan known about the role of sugar in the ciety for Nutrition, Canadian Nutrition Society, MB. A trial of sugar-free or sugar-sweetened development of obesity and diabetes. Calorie Control Council, Diabetes and Nutrition beverages and body weight in children. N Engl Study Group of the European Association for the – Although no one would take issue that J Med 2012;367:1397 1406 Study of Diabetes, International Life Sciences 16. Ebbeling CB, Feldman HA, Chomitz VR, et al. we are indeed in the midst of an obesity Institute North America and Brazil, Pulse Can- A randomized trial of sugar-sweetened bever- and diabetes epidemic, placing the ada, Dr. Pepper Snapple Group, and The Coca- ages and adolescent body weight. N Engl J Med fi blame on sugar consumption lacks per- Cola Company. He is also an unpaid scienti c 2012;367:1407–1416 advisor for the International Life Sciences Insti- suasive evidence and is misguided. Al- 17. Hu FB, Malik VS. Sugar-sweetened bever- tute North America, Food, Nutrition, and Safety ages and risk of obesity and type 2 diabetes: though calories from sugar (sucrose, Program and spouse of an employee of Unilever epidemiologic evidence. Physiol Behav 2010; d Canada. No other potential conflicts of interest fructose, or HFCS in any form solid or 100:47–54 relevant to this article were reported. liquid) have been shown to increase 18. Mozaffarian D, Hao T, Rimm EB, Willett WC, weight in a hypercaloric diet and de- Hu FB. Changes in diet and lifestyle and long- crease weight in a hypocaloric diet, References term weight gain in women and men. N Engl J Med 2011;364:2392–2404 when consumption is corrected for en- 1. White JS. Straight talk about high-fructose 19. Schoeller DA, Thomas D, Archer E, et al. ergy intake, sugar has no effect on body : what it is and what it ain’t. Am J Self-report-based estimates of energy intake Clin Nutr 2008;88:1716S–1721S weight. Finally, there is no direct evi- offer an inadequate basis for scientificcon- 2. Sun SZ, Empie MW. Fructose metabolism in dence that sugar itself, in liquid or solid clusions. Am J Clin Nutr 2013;97:1413–1415 humansdwhat isotopic tracer studies tell us. 20. Subar AF, Thompson FE, Kipnis V, et al. form, causes an increase in appetite, Nutr Metab (Lond) 2012;9:89 decreases satiety, or causes diabetes. 3. Rippe JM, Angelopoulos TJ. Sucrose, high- Comparative validation of the Block, Willett, If there are any adverse effects of fructose corn syrup, and fructose, their metab- and National Cancer Institute food frequency questionnaires: the Eating at America’sTable sugar, they are due entirely to the calo- olism and potential health effects: what do we – Study. Am J Epidemiol 2001;154:1089–1099 ries it provides, and it is therefore indis- really know? Adv Nutr 2013;4:236 245 4. Bray GA, Nielsen SJ, Popkin BM. Consump- 21. Schoeller DA, Bandini LG, Dietz WH. Inac- fi tinguishable from any other caloric tion of high-fructose corn syrup in beverages curacies in self-reported intake identi ed by food. Excess total energy consumption may play a role in the epidemic of obesity [pub- comparison with the doubly labelled water seems far more likely to be the cause of lished correction appears in Am J Clin Nutr. method. Can J Physiol Pharmacol 1990;68: – obesity and diabetes. Although many in- 2004;80:1090]. Am J Clin Nutr 2004;79:537–543 941 949 22. Heymsfield SB, Darby PC, Muhlheim LS, dividuals can lose a substantial amount 5. Bray GA. Fructose: pure, white, and deadly? Fructose, by any other name, is a health hazard. Gallagher D, Wolper C, Allison DB. The calorie: of weight and thereby also delay the J Diabetes Sci Tech 2010;4:1003–1007 myth, measurement, and reality. Am J Clin Nutr onset of diabetes, to do so has relied 6. Bray GA. Energy and fructose from bever- 1995;62(Suppl.):1034S–1041S on an overall reduction in energy con- ages sweetened with sugar or high-fructose 23. Martin LJ, Su W, Jones PJ, Lockwood GA, sumption. Thus, if reduced energy in- corn syrup pose a health risk for some people. Tritchler DL, Boyd NF. Comparison of energy Adv Nutr 2013;4:220–225 intakes determined by food records and doubly take is desirable, all caloric foods are 7. Bray GA, Popkin BM. Calorie-sweetened bev- labeled water in women participating in a candidates. A reduction in consump- erages and fructose: what have we learned 10 dietary-intervention trial. Am J Clin Nutr 1996; tion of added sugars should head the years later. Pediatr Obes 2013;8:242–248 63:483–490 962 Point-Counterpoint Diabetes Care Volume 37, April 2014

24. Heitmann BL, Lissner L, Osler M. Do we eat and risk of metabolic syndrome and type 2 di- affecting insulin sensitivity or ectopic in less fat, or just report so? Int J Obes Relat Metab abetes: a meta-analysis. Diabetes Care 2010;33: healthy humans. Am J Clin Nutr 2006;84:1374– Disord 2000;24:435–442 2477–2483 1379 25. Trabulsi J, Schoeller DA. Evaluation of die- 41. Janket SJ, Manson JE, Sesso H, Buring JE, 55. Silbernagel G, Machann J, Unmuth S, et al. tary assessment instruments against doubly la- Liu S. A prospective study of sugar intake and Effects of 4-week very-high-fructose/glucose beled water, a biomarker of habitual energy risk of type 2 diabetes in women. Diabetes Care diets on insulin sensitivity, visceral fat and intake. Am J Physiol Endocrinol Metab 2001; 2003;26:1008–1015 intrahepatic lipids: an exploratory trial. Br J 281:E891–E899 42. de Koning L, Malik VS, Rimm EB, Willett WC, Nutr 2011;106:79–86 26. Bandini LG, Schoeller DA, Cyr HN, Dietz WH. Hu FB. Sugar-sweetened and artificially sweetened 56. Faeh D, Minehira K, Schwarz JM, Periasamy Validity of reported energy intake in obese and beverage consumption and risk of type 2 dia- R, Park S, Tappy L. Effect of fructose overfeeding nonobese adolescents. Am J Clin Nutr 1990;52: betes in men. Am J Clin Nutr 2011;93:1321– and fish oil administration on hepatic de novo 421–425 1327 lipogenesis and insulin sensitivity in healthy 27. Heitmann BL. The influence of fatness, 43. Fagherazzi G, Vilier A, Saes Sartorelli D, men [published correction appears in Diabetes weight change, slimming history and other life- Lajous M, Balkau B, Clavel-Chapelon F. Con- 2006;55:563]. Diabetes 2005;54:1907–1913 style variables on diet reporting in Danish men sumption of artificially and sugar-sweetened 57. Stanhope KL, Schwarz JM, Keim NL, et al. and women aged 35–65 years. Int J Obes Relat beverages and incident type 2 diabetes in the Consuming fructose-sweetened, not glucose- Metab Disord 1993;17:329–336 Etude Epidemiologique aupres des femmes de sweetened, beverages increases visceral adipos- 28. Heitmann BL, Lissner L. Dietary underre- la Mutuelle Generale de l’Education Nationale- ity and lipids and decreases insulin sensitivity in d fi porting by obese individuals is it speci cor European Prospective Investigation into Cancer overweight/obese humans. J Clin Invest 2009; fi – non-speci c? BMJ 1995;311:986 989 and Nutrition cohort. Am J Clin Nutr 2013;97: 119:1322–1334 – 29. Sawaya AL, Tucker K, Tsay R, et al. Evalua- 517 523 58.CozmaAI,SievenpiperJL,deSouzaRJ,etal. ’ tion of four methods for determining energy 44. Hodge AM, English DR, O Dea K, Giles GG. Effect of fructose on glycemic control in diabetes: fi intake in young and older women: comparison Glycemic index and dietary ber and the risk of a systematic review and meta-analysis of controlled – with doubly labeled water measurements of to- type 2 diabetes. Diabetes Care 2004;27:2701 feeding trials. Diabetes Care 2012;35:1611–1620 tal energy expenditure. Am J Clin Nutr 1996;63: 2706 59. Sievenpiper JL, Chiavaroli L, de Souza RJ, – 491 499 45. Meyer KA, Kushi LH, Jacobs DR Jr, Slavin J, et al. ‘Catalytic’ doses of fructose may benefit 30. Black AE, Bingham SA, Johansson G, Coward Sellers TA, Folsom AR. Carbohydrates, dietary glycaemic control without harming cardiometa- fi WA. Validation of dietary intakes of and ber, and incident type 2 diabetes in older bolic risk factors: a small meta-analysis of ran- energy against 24 hour urinary N and DLW en- women. Am J Clin Nutr 2000;71:921–930 domised controlled feeding trials. Br J Nutr ergy expenditure in middle-aged women, retired 46. Colditz GA, MansonJE,StampferMJ, 2012;108:418–423 men and post-obese subjects: comparisons with Rosner B, Willett WC, Speizer FE. Diet and risk 60. Johnson RJ, Nakagawa T, Sanchez-Lozada LG, validation against presumed energy require- of clinical diabetes in women. Am J Clin Nutr et al. Sugar, uric acid, and the etiology of diabetes ments. Eur J Clin Nutr 1997;51:405–413 1992;55:1018–1023 and obesity. Diabetes 2013;62:3307–3315 31. Hu FB. Dietary pattern analysis: a new di- 47. Melanson KJ, Zukley L, Lowndes J, Nguyen 61. Dhingra R, Sullivan L, Jacques PF, et al. Soft rection in nutritional epidemiology. Curr Opin V, Angelopoulos TJ, Rippe JM. Effects of high- drink consumption and risk of developing cardio- Lipidol 2002;13:3–9 fructose corn syrup and sucrose consumption metabolic risk factors and the metabolic syn- 32. Randall E, Marshall JR, Graham S, Brasure J. on circulating glucose, insulin, leptin, and ghre- drome in middle-aged adults in the community. Patterns in food use and their associations with lin and on appetite in normal-weight women. Circulation 2007;116:480–488 nutrient intakes. Am J Clin Nutr 1990;52:739–745 Nutrition 2007;23:103–112 62. Johnston RD, Stephenson MC, Crossland H, 33. Fung TT, Rimm EB, Spiegelman D, et al. As- 48. Teff KL, Elliott SS, Tschop¨ M, et al. Dietary et al. No difference between high-fructose and sociation between dietary patterns and plasma fructose reduces circulating insulin and leptin, biomarkers of obesity and cardiovascular dis- attenuates postprandial suppression of ghrelin, high-glucose diets on liver triacylglycerol or bio- ease risk. Am J Clin Nutr 2001;73:61–67 and increases triglycerides in women. J Clin En- chemistry in healthy overweight men. Gastro- – 34. DellaValle DM, Roe LS, Rolls BJ. Does the docrinol Metab 2004;89:2963–2972 enterology 2013;145:1016 1025.e2 consumption of caloric and non-caloric bever- 49. Aeberli I, Gerber PA, Hochuli M, et al. Low 63. Sievenpiper JL; Toronto 3D (Diet, Digestive ages with a meal affect energy intake? Appetite to moderate sugar-sweetened beverage con- Tract, and Disease) Knowledge Synthesis and 2005;44:187–193 sumption impairs glucose and metabolism Clinical Trials Unit. Fructose: where does the – 35. Almiron-Roig E, Palla L, Guest K, et al. Fac- and promotes inflammation in healthy young truth lie? J Am Coll Nutr 2012;31:149 151 tors that determine energy compensation: a sys- men: a randomized controlled trial. Am J Clin 64. Flegal KM, Carroll MD, Kuczmarski RJ, tematic review of preload studies. Nutr Rev Nutr 2011;94:479–485 Johnson CL. Overweight and obesity in the United – 2013;71:458–473 50. Maersk M, Belza A, Stødkilde-Jørgensen H, States: prevalence and trends, 1960 1994. Int J 36. DiMeglio DP, Mattes RD. Liquid versus solid et al. Sucrose-sweetened beverages increase fat Obes Relat Metab Disord 1998;22:39–47 : effects on food intake and body storage in the liver, muscle, and visceral fat de- 65. Flegal KM, Carroll MD, Ogden CL, Johnson weight. Int J Obes Relat Metab Disord 2000;24: pot: a 6-mo randomized intervention study. Am CL. Prevalence and trends in obesity among US 794–800 J Clin Nutr 2012;95:283–289 adults, 1999-2000. JAMA 2002;288:1723–1727 37. Houchins JA, Burgess JR, Campbell WW, 51. Aeberli I, Hochuli M, Gerber PA, et al. Mod- 66. Flegal KM, Carroll MD, Ogden CL, Curtin LR. et al. Beverage vs. solid and vegetables: erate amounts of fructose consumption impair Prevalence and trends in obesity among US effects on energy intake and body weight. Obe- insulin sensitivity in healthy young men: a ran- adults, 1999-2008. JAMA 2010;303:235–241 sity (Silver Spring) 2012;20:1844–1850 domized controlled trial. Diabetes Care 2013; 67. Flegal KM, Carroll MD, Kit BK, Ogden CL. 38. Livesey G, Taylor R. Fructose consumption 36:150–156 Prevalence of obesity and trends in the distri- and consequences for glycation, plasma triacyl- 52. Stanhope KL, Griffen SC, Bremer AA, et al. bution of among US adults, glycerol, and body weight: meta-analyses and Metabolic responses to prolonged consumption 1999-2010. JAMA 2012;307:491–497 meta-regression models of intervention stud- of glucose- and fructose-sweetened beverages 68. Popkin BM, Nielsen SJ. The sweetening of ies. Am J Clin Nutr 2008;88:1419–1437 are not associated with postprandial or 24-h the world’s diet. Obes Res 2003;11:1325–1332 39. de Souza RJ, Bray GA, Carey VJ, et al. Effects glucose and insulin excursions. Am J Clin Nutr 69. Welsh JA, Sharma AJ, Grellinger L, Vos MB. of 4 weight-loss diets differing in fat, protein, 2011;94:112–119 Consumption of added sugars is decreasing in the and carbohydrate on fat mass, lean mass, vis- 53. Moore MC, Davis SN, Mann SL, Cherrington United States. Am J Clin Nutr 2011;94:726–734 ceral adipose tissue, and hepatic fat: results AD. Acute fructose administration improves 70. U.S. Department of Agriculture. Economic from the POUNDS LOST trial. Am J Clin Nutr oral glucose tolerance in adults with type 2 di- research service. Available from http://www. 2012;95:614–625 abetes. Diabetes Care 2001;24:1882–1887 ers.usda.gov/data-products/food-availability- 40. Malik VS, Popkin BM, Bray GA, Despres´ JP, 54. Leˆ KA, Faeh D, Stettler R, et al. A 4-wk high- %28per-capita%29-data-system.aspx#. Willett WC, Hu FB. Sugar-sweetened beverages fructose diet alters lipid metabolism without Ue1bGG0Zsz4. Accessed 21 October 2013