Study on Carbohydrates

What little work has been done in humans does little to answer the question. Two correlative comparisons ofusers and nonusers of artificial sweeteners showed that the sweeteners had no effect on body weight (18, 19). In contrast, an epidemiological study of 78 694 women found that reported weight gain was greater in those who used artificial sweeteners than in those who did not (20). There are only three published studies that have used a causative amount when approach. ofweight ate APM replaced In one, dieters who two, during were either hether artificial sweeteners aid intake and body weight, we gave free-living, normal-weight subjects 1 150 g soda sweetened with aspartame (APM) or high-fructose corn syrup (HFCS) per day. Relative to when no soda was given, drinking APM-sweetcontrol of long-term food ened soda for 3 wk significantly reduced calorie intake drinking the of both females weight sweetened take (n = 9) and males (n = of males but not of females. soda body for 3 wk significantly and 2 1) and However, decreased the body HFCScalorie in- increased

Downloaded from www. ajcn. org by guest on June 1, 2011 eight of both sexes. Ingesting either type of soda reduced intake of sugar from the diet without affecting intake of other nutrients. Drinking large volumes of APMsweetened soda, in contrast to drinking HFCS-sweetened soda, reduces sugar intake and thus may facilitate the control of calorie intake and body weight. Am J Gun Nutr 1990; 5 1:963-9. encouraged lost the same or discouraged and to use APM-sweetened (2 1). In the other fewer calories all sucrose products hospitala 6- or ized lean period obese subjects KEY WORDS tose corn syrup, Human sugar, food intake, aspartame, body high-fruc- weetness, weight, weight control Introduction It is generally benefit believed taste that artificial sweeteners (1). provide Indeed, the foods of a desirable without calories and drinks containing these substances are frequently labeled “diet. ” However, the possibility that sweet, low-calorie foods and drinks actually lead to a reduction in body weight has not been examined in detail. There is mounting evidence that in the short term (< 12 h), consumption of artificial sweeteners increases the motivation to eat. Rats increase food intake after drinking a saccharin solution (2).

Humans report increased hunger after drinking solu- than when they were fed a high-sucrose diet (22, 23). None of the work to date has examined the effect on food intake or body weight ofadding artificial sweeteners to the normal diet. In the present study, we attempted to do this by determining the effect on long-term (3-wk) food intake and body weight of consuming APM given in soda, the most prevalent vehicle for artificial sweeteners. By comparing periods when subjects drank APM, HFCS, and no soda, we planned to examine the effect of APM both as an addition to the diet and as a l2-d sugar substitute. n the diet Methods Recruitment of subjects tions Food than These of aspartame (APM), saccharin, or acesulfame-K (3, 4). intake is greater after eating a saccharin-sweetened yogurt after a glucose-sweetened or unsweetened yogurt (5). results are not caused by a postingestive or pharmacolog- The experiment was run in two replications, held in the fall of 1987 and the spring of 1988. It was approved by the Cornmittee on Studies Involving Human Beings at the University of Pennsylvania. Potential subjects were first attracted by advertisements I 2 cal effect of the artificial sweeteners; rats eat more food after sham-drinking (ingesting but not absorbing) sucrose solution (6), and humans increase hunger ratings after chewing a gum base sweetened with as little as 0. 6 mg APM (7). Moreover, subjects who have normal sweetness perception while drinking a sweet milk shake subsequently eat more food than do subjects who cannot perceive the milk shake as sweet [because of treatment with gymnemic acid (8)]. These and other findings (9) posted the Monell on local university campuses. Upon ar- From Chemical Senses Center, Philadelphia. Competitive Re- Supported y the US Department ofAgriculture’s search Grants Program grant 87-CRCR- 1-2316. 3Address reprint requests to MG Tordoff, Monell Chemical Center, 3500 Market Street, Philadelphia, PA 19104. Senses Received Accepted May 30, 1989. for publication August 9, 1989. Am iC/in Nutr 1990;51:963-9. Printed in USA. © 1990 American Society for Clinical Nutrition 963 964 TABLE 1 TORDOFF AND ALLEVA each subject was weighed (wearing casual clothes, to the nearest 100 g; the weight was not revealed to the subject), the dietary record from the previous week was examined for ambiguities, and printed instructions for the following week were given.

In the two soda conditions, subjects were directed to drink four sodas a day, keep unopened bottles in a refrigerator, and record the time each bottle they were notified, was consumed. In the no-soda “There are no special instructions condition, for this Constituents of aspartame-sweetened (APM) and high-fructose-cornsyrup-sweetened (HFCS) sodas ingested daily during 21-d test periods Constituent APM HFCS Weight(g) Water(mL) 1135 1130 1135 1000 APM(mg) HFCS(g) Calories (kcal) 590 1 3 0 133 530 week. ” bottles carrying collected At the end of the weekly of soda for the following the sodas was somewhat them in smaller batches isit, subjects were given 28 week (if necessary). Because cumbersome, a few subjects more frequently. rival at the laboratory tive subject received for an initial a written appointment, description each of the prospecstudy and Debriefing and taste tests signed a participation consent form. The study’s purpose was stated as “an ongoing project to examine basic mechanisms of food preference, food intake, and appetite. ” The only procedural details given were the requirement to keep a dietary record and “you will receive beverages to drink on various days,” but “we you will receive cannot tell you at this or what they contain. time how many The description drinks also included notice ofthe requirement to attend a weekly interview at the laboratory and a schedule of remuneration, totalling $ 100 for satisfactory completion ofthe experiment. Subjects were administered the 40-question eating attitudes test (EAT-40) (24), the 5 1-question Restrained Eating Questionnaire (25), and other questionnaires to assess medical history, food preferences, eating attitudes, and dietary restraint. On the basis of questionnaire responses, applicants were excluded ifthey were recently or currently dieting, were avoiding caffeine, had a family history ofdiabetes, or were pregnant.

Initial training period At the end ofthe 9-wk test period, taste tests were conducted to see if subjects could recognize differences between soda contaming APM and HFCS. First, each subject received a series of 16 counterbalanced triangle tests: the subject attempted to pick the disparate soda from three 10-mL samples of soda, two of one variety and one of the other. Second, the subject was allowed to drink as much as he or she wanted from four cups of soda. He or she was asked to identify whether the soda was a diet or regular type.

Unbeknownst to the subject, two glasses contained APM-sweetened soda and two, HFCS-sweetened soda. Finally, we asked what the subject thought the study was about. Analysis ofdietary records Downloaded from www. ajcn. org by guest on June 1, 2011 Dietary diet-analysis records software were analyzed (release 3. 0, by use of NUTRITIONIST-3 N-Squared Computing, Sil- An experienced registered dietitian instructed each subject on how to complete dietary records. The 45-mm lesson emphasized the necessity of timely and accurate record keeping and included demonstrations with food models and household measures.

To augment compliance, subjects were told, “We could determine what you have eaten from analysis of urine samples” (although this was untrue). To ensure understanding ofthe instructions, subjects kept a practice dietary record for 2 or 3 d. The completed record was scrutinized by the dietitian (with the subject present) to clarify any ambiguities and to familiarize subjects with the rigor required for keeping a dietary record. At this stage six females and eight males elected to quit the experiment. Two males who kept insufficiently detailed records were also eliminated. Experiment design and procedure erton, OR) by trained personnel who were unaware of the treatment conditions. Components of foods not listed in the database were obtained directly from the manufacturers or by chemical analysis. For simplicity, we combined fructose, glucose, sucrose, and other monoand disaccharides as “sugar. ” After inspection of initial results, separate values were derived for sugar in beverages (ie, soft drinks, coffee, and tea) and food (all other sources of sugar). Results Preliminary analyses found there were no differences be- Each subject maintained a dietary record continuously for 9 wk. During this eriod they received, in counterbalanced order, for 3 wk each, soda sweetened with APM, soda sweetened with high-fructose corn syrup (HFCS), or no experimental drinks. The cola-flavored soda was provided in ‘ 300-mL glass bottles. There was an alphanumeric code on the cap or sleeve ofeach bottle but nothing to inform the subject ofthe identity of the drink. During the appropriate periods, subjects were required to drink four bottles ( 1 1 35 g) ofsoda daily (Table 1). At the start of the test period and then at weekly intervals, tween the results of the two replications of this study, so they were combined.

Of the I 3 female and 28 male subjects who started the study, 1 female and 5 males stopped keeping dietary records or failed to keep appointments at the laboratory. Three females were eliminated because of chicken pox, pneumonia, and relocation away from the area. Two males complained about having to drink so much soda, so they were also dropped from the study. Analyses and data presentation are based on the remaining 9 females and 2 1 males. Subject characteristics Anthropometric measures are shown in Table 2. Body mass indexes ofthe females and males were 25. 4 ± 1. 4 and 25. 1 ± 0. kg/m2, respectively, which falljust below the 75th percentile of body weight distribution (26). With the exception offour males who ate fixed meals four times per week, all subjects controlled their own food choice and meal size. There were minimal re- ARTIFICIAL TABLE Physical SWEETENERS AND BODY WEIGHT 965 2 characteristics of subjects* Female (n 9) Characteristic Age (y) Height (cm) = ± ± ± (n Male = 21) ± ± ± 28. 2 165. 5 2. 7 2. 2 4. 3 22. 9 174. 5 76. 6 0. 8 1. 2 Weight (kg) *j±5EM 69. 6 2. 1 ports of food allergies or aversions. scores on the EAT-40, a measure 1 1 . 8 ± 2. , males 9. 9 ± 1. 1). The ing Questionnaire revealed normal No subjects had extreme of eating disorders (females 5 1-question Restrained eating behavior except Eatthat two females and one male had high (> 2 SD above the mean) restraint (factor 1) scores and five males had high disinhibition (factor 2) scores. None of the questionnaire responses corre- lated ofthe significantly except Restrained with Eating food intake Questionnaire) or weight between and = change hunger calorie p < during (factor intake the 3 dur- experiment, for a correlation ing the no-soda Body weight baseline) period (r 0. 37, 0. 05). Subjects gained slightly but wk of drinking HFCS-sweetened significantly more weight after 2 soda than after the same pesoda or no experimental soda was more marked after 3 wk (Fig did males durHFCS-sweet- riod drinking APM-sweetened (Appendix A). This difference 1). Females lost significantly more weight than ing the control (no-soda) period. While drinking ened soda, females gained p < 0. 0 1) and males gained drinking APM-sweetened ± 0. 29 kg, p kg, < weight significantly (0. 97 ± 0. 25 kg, slightly (0. 52 ± 0. 23 kg, NS).

While soda, females lost gained weight slightly (0. 47 males weight significantly (0. 25 ± 0. 22 NS) but 0. 05). Thus, the effect on both sexes combined and days (1-21 d). Separate analyses were performed either including or excluding the ingredients from the experimental sodas. All the analyses found that females consumed significantly less than did males, and there was no interaction between sex and treatment (Appendix A). None ofthe analyses produced a main effect or interaction involving the days factor, indicating that intakes were stable across the 2 l-d treatment periods.

The possibility ofcarry-over effects from one period to another was examined using the same procedure as for body weight data. Results from the first 3-wk period were analyzed separately by using between-subject comparisons (Appendix B). The results of these analyses from a period before carry-over effects could have occurred were similar to those from the complete set of data, indicating that carry-over effects were either absent or, if present, undetectable and thus ofminor significance. Calories.

Relative to calorie intake during the no-soda condition, drinking 530 kcal HFCS-sweetened soda/d produced a large and highly significant increase in total calorie intake (including calories in the experimental soda). Drinking the same volume ofAPM-sweetened soda decreased calorie intake. Both APM and HFCS consumption significantly reduced intake of calories from the diet (ie, calories excluding the sodas) to the same extent (by 179 and 195 kcal/d, respectively; Table 3). The decrease in dietary calorie intake produced by drinking either form of soda was due entirely to a decrease in sugar intake (Fig 2).

Drinking soda did not affect the intake of protein, fat, alcohol, or complex (nonsugar) carbohydrate (Table 3). Sugar andsoda. During the period without experimental sodas, average intake of sugar-sweetened soda was 292 ± 1 33 g for females and 414 ± 85 g for males. Three females and two males drank essentially no (< 25 g/d) HFCS-sweetened soda; one female and two males drank > 1 135 g/d. Intake of APMsweetened soda during the same period was 1 59 ± 82 g for females and 88 ± 40 g for males, which included 6 females and 16 males who did not drink any. The total intake ofboth types

Downloaded from www. ajcn. org by guest on June 1, 2011 ofdrinking HFCS-sweetened body weight, whereas the soda crease was to nonsignificantly in weight seen when gain soda was to significantly increase effect of drinking APM-sweetened decrease males it. Female =9) Male (n=2 1) Because APM could of the counterbalanced reflect either a direct caused body design drank of the study, soda sweetened period the dewith influence by a previous ofthe soda or recovof HFCS- ery from possibilities, the weight we sweetened-soda consumption. compared To discriminate weight changes between of the these three ) C male and three female subgroups of subjects during the first 3 wk of the experiment (Appendix B) and during each of the three 3-wk periods of the study (Appendix C). The pattern of results for each of the periods was more-or-less similar to that seen overall, although because of the smaller group loss in sensitivity produced by the use ofbetween-subject parisons, the only significant during the first 3-wk period difference for females 0 -C 0 .4. J ii No -1 sizes and corn- ci) was present >‘ 0 0 (Appendix who effects loss drank seen C). Judging APM-sweetened weight when per se. ofbody y the desoda gain could drank FIG Soda APM crease in the weight before any possible occur, soda it appears containing that of males carry-over the weight subjects HFCS APM was due to the soda Food intake and of total calories were anwith factors of sex, treatment, Intakes of the various nutrients alyzed by three-way ANOVAs in body weight during 3-wk periods when subjects sweetened with aspartame (APM), an equal weight ofsoda sweetened with high-fructose corn syrup(HFCS), or had no experimental manipulation (no soda). *p < 0. 05 relative to weight gain in no-soda period. 1. Changes drank 150 g/d of soda 966 TABLE 3 ofdrinking Effect on dietary nutrient intake sweetened with APM or HFCS* Measure Nosoda TORDOFF AND ALLEVA evidence that the effects weight were influenced soda. ofthe sodas on calorie intake and body by the subjects’ ability to identify the guessed the experiment’s purpose; most soda (1 135 g/d) APM kcal/d HFCS None ofthe subjects Females(n = 9) Alcohol Fat Protein Carbohydrate 65±23 747±93 266±25 39±12 745±95 262±28 58±23 726± 84 256± 27 405 ± thought we were performing market a new brand of soda. None noticed changed their body weight or altered take or selection.

Discussion research of some kind on that drinking the sodas their patterns of food in- Imposing the requirement to drink 1 135 g/d of APM-sweetened soda on normal-weight, freely feeding subjects decreased calorie intake significantly (by 7%) and reduced body weight slightly (significantly in males). This was in marked contrast to 261±60 255±55 225±56 the highly significant, 13% increase in calorie intake and sig932±45 945±56 937 ±41 nificant increase in body weight produced by consumption of 373±23 384±27 373± 18 the same amount of HFCS-sweetened soda.

The two types of soda produced an identical, 33% decrease in dietary sugar in617±43 612 ±48 ComplexCHO 624±49 461 ± 38t take (excluding the sugar in the soda), without affecting intake Sugar 674 ± 49 453 ± 40t Totalintake 2801 ± 150 2647± 153 2645± l24 t ofother macronutrients. This was caused in part by the “experimental” sodas displacing discretionary beverages; subjects i ±SEM. given four bottles ofsoda per day have little motivation to purt Significantly different from no-soda condition: tp < 0. 005, f#{231}p chase and drink their own. However, drinking either form of