Soft drinks with aspartame: Effect on subjective hunger, food selection, and food intake of young adult males

Physiology&Behavior,Vol. 49, pp. 803--810. © Pergamon Press plc, 1991. Printed in the U.S.A.

0031-9384/91 $3.00 + .00

Soft Drinks With Aspartame: Effect on Subjective Hunger, Food Selection, and Food Intake of Young Adult Males RICHARD M. BLACK,*t

PHYLLIS TANAKA,*

LAWRENCE

A. L E I T E R * t

A N D G. H A R V E Y A N D E R S O N .1

*Department of Nutritional Sciences, #Department of Medicine, "~Department of Psychiatry Faculty of Medicine, University of Toronto, Toronto, Canada R e c e i v e d 15 M a y 1990 BLACK, R. M., P. TANAKA, L. A. LEITER AND G. H. ANDERSON. Soft drinks with aspartame: Effect on subjective hunger, food selection, andfood intake of young adult males. PHYSIOL BEHAV 49(4) 803-810, 1991.--Ingestion of aspartame-sweetened beverages has been reported to increase subjective measures of appetite. This study examined the effects of familiar carbonated soft drinks sweetened with aspartame on subjective hunger, energy intake and macronutrient selection at a lunch-time meal. Subjects were 20 normal weight young adult males, classified as either restrained or nonrestralned eaters. Four treatments of carbonated beverages included 280 ml of mineral water, one can of a soft drink (280 ml) consumed in either 2 or 10 minutes, or two cans of a soft drink (560 ml) consumed in 10 minutes, administered at 11:00 a.m. Subjective hunger and food appeal were measured from 9:30 a.m. to 12:30 p.m., and food intake data were obtained from a buffet lunch given at 12:00 noon. There were no treatment effects on energy intake, macronutrient selection or food choice at the lunch-time meal, or food appeal, though restrained eaters consumed more than nonrestralned eaters in all four treatment conditions. Consumption of two soft drinks (560 ml, 320 mg aspartame) significantly reduced subjective hunger from 11:05 a.m. to 11:30 a.m. compared to one soft drink (280 ml, 160 mg aspartame) or 280 ml of mineral water. Thus ingestion of soft drinks containing aspartame did not increase short-term subjective hunger or food intake. Aspartame

Subjective hunger

Human behavior

Feeding

THE nonnutritive sweetener aspartame (APM) is becoming an increasingly popular sucrose substitute in foods and beverages. However, when such high intensity sweeteners are used as replacement for nutritive sweeteners in foods and beverages, the effect on food intake is uncertain. There are various reports suggesting that such replacement may reduce (13, 25, 33), have no effect (8,15) or even enhance food intake (4, 20, 21). The possibility of enhancing food intake through consumption of nonnutritive sweeteners was raised by reports of an increase in subjective hunger following sweetener ingestion in water [e.g., (4,21)], in yoghurt (20), as the sole sweetener in unflavored gum (30) and by the correlation of nonnutritive sweetener usage with increased rate of weight gain in an epidemiological study (28). The putative mechanism for the appetite-enhancing effect of nonnutritive sweeteners has been elicitation of a cephalic response resulting from the taste of the sweetener (20,30). Cephalic responses can stimulate insulin release, increasing glucose uptake by tissue, which in turn reduces blood glucose levels and ostensibly stimulates appetite [cf. (14, 17, 18, 32)]. It has been hypothesized that, in contrast to nutritive sweeteners, nonnutritive

sweeteners elicit this cephalic response without providing energy, and so the stimulated appetite is not assuaged (20,32). In addition to increases in subjective hunger, increased food intake arising from the consumption of a food with added nonnutritive sweeteners has also been reported. One hour after ingestion of a saccharin-sweetened yoghurt preload, subjective ratings of hunger increased as did lunch-time intake and total calories consumed during the remainder of the day, compared to consumption of an unsweetened yoghurt preload (20). However, others have failed to find any effect of the addition of nonnutritive sweeteners to food on energy consumption, for example when added to a breakfast cereal (15), or pudding or jello (25). It has been suggested that this failure to observe any appetite enhancement after consuming nonnutritive sweeteners may have been due in part to the use of foods as the vehicle rather than beverages [as in the initial positive finding (4)] (26). However, both short-term (18,26) and long-term (31) studies have failed to uncover any appetite or food intake enhancement associated with consumption of nonnutritive sweeteners in flavored beverages. In the previous experimental studies involving beverage con-

1Requests for reprints should be addressed to Dr. G. H. Anderson, Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, 150 College St., Toronto, Ontario, Canada M5S 1A8.

803

804

BLACK ET AL.

sumption and the acute response on appetite and food intake, the nonnutritive sweetener has been provided as a beverage or snack, generally 30 to 60 minutes prior to presentation of a lunch-time meal. It is important to note that, without exception, these studies did not control for time since, macronutrient content of, or energy content of, the previous meal, i.e., breakfast. As such, subtle effects on appetite engendered by the nonnutritive sweetener may have been masked. Furthermore, while some have attempted to relate their f'mdings to consumption of commercially available soft drinks containing nonnutritive sweeteners such as aspartame [cf. (15, 19, 26)], no direct tests of the effects on appetite of consuming such soft drinks have been conducted. Therefore, the present study strictly controlled timing and size of the breakfast meal on test days, and delivered the nonnutritive sweetener aspartame in a commercially available soft drink, commonly referred to as a diet soft drink. We report herein the effects of both volume and duration of exposure to diet soft drinks on subjective measures of hunger and on lunch-time food intake.

TABLE 1 EXPERIMENTAL SCHEDULE Time

Procedure

10:00 p.m. (night before) 7:45 a.m. 8:00 a.m. 9:30 a.m. 10:55 a.m. 11:00 a.m.

11:05 a.m. l h l 0 a.m. 11:15 a.m. 11:20 a.m. 11:30 a.m.

METHOD 11:45 a.m.

Subjects

12:00 p.m.

Twenty male subjects [ n = 2 0 , age 19 to 25 years, Body Mass Index (BMI, weight/height 2, units are kg/m 2) 22 to 29] participated in the study. Subjects were recruited by posters placed throughout the University of Toronto and by word of mouth. Excluded from the study were those who smoked, followed a vegetarian diet, or had a diagnosis of phenylketonuria, a food allergy, a major medical illness, or received medications. The subjects gave informed consent to participate in the study, approved by the University of Toronto Human Subjects Review Committee.

12:05 p.m. 12:30 p.m.

Procedure Upon recruitment, subject weight and height were measured. Each subject also completed the Restrained Eaters Test (9), as well as a Food Liking Checklist and a Food Acceptability List, from which an individualized buffet lunch menu was prepared. Subjects also indicated a preferred soft drink flavour from a choice of Sprite, 7-UP, orange, ginger ale or cream soda, to be consumed during the study. The full experimental schedule is described in Table 1. Subjects were required to fast from 10:00 p.m. the previous night until breakfast was provided. Subjects arrived at the feeding studies laboratory in the Department of Nutritional Sciences, University of Toronto, at 7:45 a.m. Each subject was then presented with a breakfast meal (Table 2) to be entirely consumed within 30 minutes. The breakfast meal was individually tailored to each subject's weight, so that it provided 25% of the average daily caloric intake, based on tabulated estimates of average energy expenditure (e.g., 3000 kcal/day for a 71-kg man of average height and age range) (7). Breakfast calories were adjusted for an individual's weight by altering the amount of cereal and milk presented for consumption. Measures of subjective appetite and food preference were taken throughout the morning (Table 1). Four treatments were administered at 1 h 0 0 a.m., with treatment order being fully randomized across subjects (Table 3). Subjects were unaware of the nature of the soft drink, i.e., regular or diet. Soft drinks of the predetermined flavor preference for the individual were served cold (5°C) in 300-ml plastic cups. Subjects were required to consume all of the beverage provided within the time allotted. Lunch was provided at 12:05 p.m. Each subject was served an individualized buffet lunch, prepared in a visually appealing man-

Subject fasts Subject arrives at Feeding Studies Laboratory Subjects consume breakfast VAS Motivation to eat VAS Food Appeal VAS Motivation to eat VAS Food Appeal Treatment (preload) VAS Motivation to eat VAS Motivation to eat VAS Motivation to eat VAS Motivation to eat VAS Motivation to eat VAS Food Appeal VAS Motivation to eat VAS Motivation to eat VAS Food Appeal Buffet lunch served VAS Motivation to eat VAS Food Appeal Beverage Consumption checklist*

*To be answered after completing final test session only. ner and served in excess, and instructed to eat until comfortably full. Subjects were allowed 30 minutes to complete their lunch. Lunch size was sufficient to ensure that no subject would consume all that was offered in total, or of any one item. While individual lunches varied in the foods that were presented, overall macronutrient content was essentially constant across lunches. For any given subject, the buffet lunch was the same at each test session. A typical lunch menu is presented in Table 4.

Dependent Variables Food intake and selection. At the end of each lunch, food consumed was measured and total calories consumed were calculated. The number of calories consumed as protein (PRO), carbohydrate (CHO) and fat (FAT) were also calculated. Macronutrient

TABLE 2 COMPOSITION OF AN AVERAGE BREAKFAST

Food

Quantity

kcal

PRO g

CHO g

FAT g

Shreddies 2% milk Oat-bananaraisin muffin Orange juice

250 ml 175 ml 2 small

209 96 330

5.1 6.4 6.0

46.1 9.3 58.0

0.4 3.7 10.0

250 ml

118

1.8

28.4

0.2

753

16.4 (8%)

141.8 (75%)

14.3 (17%)

Optional coffee was available with a maximum addition of 10 g sugar and 30 ml milk = 50 kcal.

SOFT DRINKS WITH ASPARTAME

805

TABLE 3

10-cm line word anchored at each end. Subjective appetite was assessed with the aid of four different VAS's which measured 1) desire to eat, 2) hunger, 3) fullness, and 4) prospective consumption (27). The food appeal VAS measured the appeal (no appeal at a l l . . , very appealing) of 32 commonly available foods. Portion sizes were equated for energy content. These foods were classified by principal macronutrient component (i.e., protein, carbohydrate, fat or low calorie, 8 foods per classification for a total of 32 foods) (21).

EXPERIMENTAL DESIGN AND ASPARTAME CONTENT OF SOFT DRINKS

Treatment A B C D

Aspartame (rag)*

Preload 280 ml Carbonated Water (control) 280 ml Diet Soft Drink 280 ml Diet Soft Drink 560 ml Diet Soft Drink

Consumption (rain)

0

2

160 to 170 160 to 170 320 to 340

2 10 10

*Precise amounts depended upon the diet pop consumed.

composition of the foods was determined by chemical analysis results obtained through the manufacturer when possible, and otherwise from food composition tables. Because the foods were of relatively constant composition (with the possible exception of the fat content of the beef), it was assumed that any minor errors in macronutrient composition and total energy content reported in the food tables would not contribute substantially to the within subject variability on test versus control days. Foods chosen at the lunch-time meal were placed in the following categories: bread, meat, dairy, condiment, or dessert. Calculation of the energy content of foods (kcal) consumed from each category allowed determination of any treatment effects on selection of these foods. Visual analogue scales. Measurement of subjective appetite and food appeal were made throughout the morning using visual analogue scales (VAS) (Table 1). Each VAS was a continuous

TABLE 4 TYPICAL BUFFET LUNCH MENU

Food or Beverage Bread: Whole Wheat Meat: Ham Turkey Bologna Dairy: Cheese (cheddar) Milk (whole) Orange juice Water Condiments: Butter Mayonnaise Mustard Lettuce Dessert: Chocolate chip cookies Oatmeal cookies Oreo cookies Shortbread cookies Sara Lee Pound Cake Sara Lee Chocolate Fudge Cake

Amount

324.0 g (10 slices) 101.5 g ( 5 slices) 108.8 g ( 5 slices) 217.6 g (10 slices) 232.4 g 250 ml 250 ml 250 ml 30 g 60 g 15 g 4 leaves 5 cookies 5 cookies 5 cookies 5 cookies 3 cakes (46 g each) 3 cakes (46 g each)

Data Analysis General. The primary analysis was conducted as follows. First, an ANOVA involving all treatment conditions was performed on the data, allowing a general statement about the treatment effects on subjective appetite and food intake (34). However, one might argue that valid comparisons existed only between treatment conditions where a single variable was altered, and that an ANOVA invoNing all four treatment conditions would be unable to detect the subtle effects of any given treatment. Therefore, the following treatment pairs, where a single variable was altered, were also analyzed: A vs. B (water control vs. APM, 2 minutes); B vs. C (duration of stimulation, 2 vs. 10 minutes); C vs. D (amount of stimulus, 160 mg APM vs. 320 mg APM). Because it has been shown that responses to food vary between individuals of high and low restraint [based upon their score on the Restrained Eaters Questionnaire (9)], a secondary analysis of the data was conducted with restrained/nonrestrained as an additional main effect. Food intake and selection. Total energy (kcal) intake data were analyzed through a one-way repeated measures ANOVA (Treatment as the main effect). Macronutrient consumption was analyzed through a two-way repeated measures ANOVA (Treatment × Macronutrient), as was food choice (Treatment x Food Type). Visual analogue scales. Treatment baseline for the VAS scores was taken 5 minutes prior to consumption of the preload. The effect of treatment and time on VAS scores was calculated by subtracting the baseline VAS from the subsequent VAS scores. Thus each subject's data are expressed relative to their individual baseline values. For VAS scores of subjective appetite, the data for each question were analyzed separately by a 2-way repeated measures ANOVA, with Treatment and Time as main effects. The questions from the food appeal VAS scores were grouped by major component (protein, carbohydrate, fat, or low calorie) and then analyzed through a three-way ANOVA (Treatment × Macronutrient × Time). All post hoc comparisons were conducted using the NewmanKeuls procedure and the Studentized range statistic, q. RESULTS

Food Intake and Selection The analysis across all treatments revealed that preload had no significant effect on total caloric intake at the lunch-time meal, F(3,57)=0.29 (Fig. 1). In addition, there was no treatment order effect on food intake, F(3,57)= 1.63. Comparisons between designated treatments failed to uncover any effect due to a) aspartame-containing diet soft drink, b) duration of stimulation, or c) amount of diet soft drink consumed, on total caloric lunch-time intake [largest F(1,18)= 0.55]. Macronntrients were not consumed in equal amounts, F(2,38) = 55.0, p<0.001, post hoc analyses showing that subjects ate more FAT than CHO or PRO (q2=3.6, p<0.05; q3 = 7.4, p<0.01, respectively) and more CHO than PRO (q2 =

806

BLACK ET AL.

2000

EEE] Water

1800

I

1600 1400 v

1 Soft Drink 2 min

1 Soft Drink 10 min 7771 2 Soft Drinks 10 min

0o i

1200 1000

60O 400 200

Total Intake

PRO

n,,n 1i,il CHO

FAT

M a c r o n u t r i e n t Selected

FIG. 1. Effect of soft drink consumption on total food intake and macronutrient selection at the lunch-time meal. Data are expressed as energy (kcal) intake and show mean values plus 1 SEM. 3.8, p<0.05). However, there was no Treatment effect, F(3,57) = 0.31, or Treatment × Macronutrient interaction, F(9,171) = 0.20 (Fig. 1). Analysis of foods chosen at lunch uncovered no effect due to Treatment, F(3,57)=0.50, or Food Category, F(4,76)= 0.58, and there was no Treatment × Food Category interaction, F(12,228) = 0.80. Analysis of treatment pairs uncovered little else. Differences in Macronutrient consumption remained in all three treatment comparisons [F(2,38) ranged from 40.6 to 50.3, order of consumption being FAT > CHO > PRO], but there were no Treatment effects [largest F(1,19) = 0.63] and no interactions [largest F(2,38) =0.34]. Similarly with food choice, there were no effects due to Treatment [largest F(1,19)=1.12], or Food Category [largest F(4,76)=0.94] and there were no interactions [largest F(4,76) = 1.19].

Subjective Appetite Treatment had a significant effect on both "Desire to Eat,"

F(3,57)=5.13, p < 0 . 0 0 5 , and " F u l l n e s s , " F(3,57)=7.16, p<0.001. There were no treatment effects on " H u n g e r , " F(3,57) = 1.76, or "Prospective Consumption," F(3,57)= 1.44. Post hoc analyses (456 d30 indicated that for both "Desire to Eat" and " F u l l n e s s , " the significant F values were entirely attributable to the effects of consuming 560 ml of diet soft drink (Treatment D): the 560 ml preload of diet pop suppressed desire to eat and increased fullness. Specifically, desire to eat was less following ingestion of the 560 ml diet soft drink preload (Treatment D) than: a) mineral water (Treatment A) at 11:05, 11:10, 11:15, 11: 20, 11:30 and 11:45 a.m. (p's<0.01); b) 280 ml of diet soft drink consumed in 2 minutes (Treatment B) at 11:10, 11:15, 11:20 and 11:30 a.m. (p's<0.05); and c) 280 ml of diet soft drink consumed in 10 min (Treatment C) at 11:05, 11:10, 11:15, 11:20 and 11:30 a.m. (p's<0.05). Similarly, fullness was greater following ingestion of the 560-ml preload than: a) the mineral water preload at 11:05, 11:10, 11:15, 11:20 and 11:30 a.m. (p's<0.01); b) 280 ml of diet soft drink consumed in 2 minutes (Treatment B) at 10, 15, 20 and 30 minutes posttreatment (p's<0.05); and c) 280 ml of diet soft drink consumed in 10 min (Treatment C) at 11:05, 11:10 and 11:15 a.m. (p's<0.01). There were no differences between any of the other treatment conditions following consumption of the beverages (see Table 5 for appropriate Newman-Keuls q values). When treatment pairs were analyzed separately, effects emerged that were not detected by the overall ANOVA (Fig. 2). There was no difference between the mineral water condition and 280 ml of diet pop consumed in 2 minutes (Treatment A vs. Treatment B) on any question [largest F(1,19)= 2.90] and no difference when duration of stimulation was the variable, i.e., 280 ml of diet pop consumed in either 2 minutes or 10 minutes (Treatment B vs. Treatment C) [largest F(1,19)= 1.26]. However, consuming 560 ml of diet pop significantly lowered desire to eat, and prospective consumption, while significantly increasing subjective fullness, compared to consuming 280 ml of diet pop (Treatment C vs. Treatment D) [question (1), F(1,19)= 4.64, p<0.05; question (2), F(1,19) = 2.55, n.s.; question (3), F(1,19) = 5.71, p < 0 . 0 3 ; ques-

TABLE 5 RESULTS OF POST HOC ANALYSES(NEWMAN-KEULS)FOR VAS MEASURESOF "DESIRE TO EAT" AND "FULLNESS"a Time of Test Administration

Desire to Eat Treatment A> D Treatment B> D Treatment C> D Fullness Treatment A< D Treatment B< D Treatment C< D

11:05 a.m.

11:10 a.m.

11:15 a.m.

11:20 a.m.

11:30 a.m.

11:45 a.m.

q6=4.9t

q17 = 10t

ql9 = 10t

q2o= 8.2t

ql2=6.2t

qll =6.1t

q3 = 2.3n~

q9 = 9.0t

q14 = 7.3t

ql4 = 4.8*

q8 = 5.0*

q4 = 2.1 ns

q7=5.0 *

q6=7.1t

q12=6.4t

q15=5.6 *

qll=5.6t

q8=3.6 "s

q6=5.4t

q13 = 12t

q17=8.8t

q19=8.5t

qls=6.9t

qs=3.6 ns

q2 = 1.2ns

q9 = 11t

qll = 6.8t

qlo = 4.6*

ql5 = 5.4"

q9 = 3.9 as

q9 = 7.2t

q5 = 8.1 t

q5 = 5.0t

ql 3 = 4.7ns

q2 = 0.8ns

q7 = 4.1 ns

~q~he Newman-Keuls statistic is qr, where r is the number of steps plus one that two means are apart on a rank ordered scale. *p<0.05; tp<0.01. ~SNonsignificant.

SOFT DRINKS WITH ASPARTAME

807

o--o Water ~.oF *--* I Soft Drink - ~ rain.

,,---o

I Soft

Drink

-

~

0--o I Soft Drink - 10 min 401 .--. ~ Soft Drinks - tO rain ~o ~

rain

"°°.o , ~ i ~ 2 ~ ~ 4 ~

oo

, p ( 0.05

~

Desire

. -~.o -4.0

-20

-~o I

I

I

,.oI

t

~

,

l

l

|

zo

z.o

•

~6

2o

6/s...A---'~,-'-~'

~/~__$--,

o.o

.< -4.0 e. ~.o[

~

Hunger

C/~

~)

~.o o.o

°:I

. ~

•~---~+~. / .~.~,

.

.

.

"

oo

•

"

--

~,~ L

oo -20

-~o ,

-40

-2.0 -4O,

.

i

-~0 '

'

!

/

I

,

i

i

q

i

~

i

Fullness

-40

-20

-,io

i

Time

of

-4o

i

Test

FIG. 2. Scores from the Visual AnalogueScales (VAS) are presented. Each column represents the comparisonof two treatments, for the four questions designedto assess subjectivehunger. For each subject, baseline is the VAS score immediatelyprior to preload consumption.Statisticallysignificant differences (p<0.05) are indicatedby *. Vertical bars represent ---1 SEM.

tion (4), F(1,19)=4.20, p<0.05].

Food Appeal Treatment had no effect on food appeal measured by VAS, F(3,57)= 1.15 (Fig. 3). There were, however, significant effects due to Time, F(3,57)=26.7, p<0.001, and Major Component (CHO, PRO, FAT or LOW calorie), F(3,57)= 5.00, p<0.001, as well as a significant Time × Major Component interaction, F(9,171) = 15.8, p<0.001. Post hoc analyses (171 dr') indicated that the appeal of all foods increased from 10:55 a.m to 12:00 noon, declining after the lunch-time meal (all p's<0.01). Furthermore, FAT foods were significantly less appealing than PRO and LOW calorie foods in all treatment conditions (p's<0.01), and significantly less appealing than CHO foods in treatment conditions B and C (280 ml of diet soft drink consumed in 2 min and 10 min, respectively) (p's<0.05). Finally, analysis of the Time × Major Component interaction revealed that PRO foods increased in appeal more rapidly than all others (p's<0.01), and CHO foods increased in appeal more rapidly than LOW calorie foods (p<0.05).

ers weighed slightly more than nonrestrained eaters ( 7 8 - 4.1 kg compared to 70---2.6 kg, mean---SEM), but this difference failed to reach significance, F(1,18) = 3.4, p = 0.077. Restrained and nonrestrained eaters consumed significantly different amounts of food at the lunch-time meal, F(1,18)=6.4, p<0.02, post hoc analyses showing that restrained eaters consumed more in all treatment conditions and for all macronutrients (see Table 6). However, there was an absence of any Group × Treatment interactions, indicating that treatment did not differentially affect the intake of restrained compared to nonrestrained eaters, i.e., the label "restrained eater" carried no predictive power with regards to the effect of treatment 60 minutes prior to eating, on either meal size or macronutrient selection [largest F(1,18) = 2.01. There were no Group differences on any of the motivation to eat questions [largest F(1,18) = 1.5], though restrained eaters found PRO foods, F(1,18)= 10.3, p<0.005, and LOW calorie foods, F(1,18) = 5.5, p<0.03, less appealing than did nonrestrained eaters. Once again, however, there were no Group x Treatment interactions [largest F(3,54)=0.3]. Analysis of treatment pairs did not uncover any additional Group effects or Group x Treatment interactions.

Restrained Eaters Subjects were classified as either restrained or nonrestrained eaters, based upon their results on the Restrained Eaters Questionnaire (9): restrained eaters scored greater than 14 (n = 7) while nonrestrained eaters scored less than 11 (n = 13). Restrained eat-

Beverage Consumption Checklist Everyday soda consumption for these subjects ranged from 0 to 4 (280 ml) cans per day, with a mode of 0 and a median of 1. Of those that regularly consumed soda pop, six of ten drank rag-

808

BLACK ET AL.

6.0

6.0

CHO foods

5.0

FAT foods

5.0

o--o

*ater

0--0

1 Soft Drink - 2 m i n

[~--[-'] I Soft Drink - I0 rain

v

4.0

40

3.0

3.0

2.0

20

1.0

10'

OO

O0

m--"

2 Soft Drinks - 10 rain

¢,.)

a~ o ~.)

6.0

PRO

50

60

foo~

4.0

LOW c a l o r i e f o o d s

50

0 iJ\

>

40

3.0

30

2.0

20

10

10 00

0.0

10:55

I 1:30

12:00

12:30

10:55

Time

of

11 3 0

12:00

12:30

Test

FIG. 3. Food appeal VAS scores for each of the four food types: predominantly carbohydrate (CHO), protein (PRO), fat (FAT) and low calorie (LOW) foods. Vertical bars represent --- 1 SEM.

ular soda, one drank diet soda, and the r e m a i n i n g three drank equal a m o u n t s o f both. Following the final test session, only three o f twenty subjects identified the flavored (aspartame) beverage as a diet soda, n o n e o f these three b e i n g diet soda drinkers. N o n e o f these m e a s u r e s correlated in any w a y with l u n c h - t i m e intake, motivation to eat, or food appeal. DISCUSSION This e x p e r i m e n t failed to u n c o v e r a n y e n h a n c e m e n t o f appetite following ingestion o f an a s p a r t a m e - s w e e t e n e d soft drink. Neither v o l u m e nor duration o f c o n s u m p t i o n o f the diet soft drink altered subjective m e a s u r e s o f appetite 60 m i n u t e s later. Similarly, the three diet soft drink treatments produced no alteration

in total energy c o n s u m e d at the l u n c h - t i m e m e a l , macronutrient selection at the l u n c h - t i m e m e a l , foods c h o s e n for c o n s u m p t i o n at the lunch-time meal, or food appeal m e a s u r e d on visual analogue scales, c o m p a r e d to the carbonated mineral water preload. Finally, as other studies have f o u n d (23), restrained eaters cons u m e d m o r e than nonrestrained in a laboratory setting. H o w e v e r , treatment did not differentially affect food intake nor motivation to eat in these g r o u p i n g s , t h o u g h it did have a marginal affect on the appeal o f h i g h P R O foods. Little data exists on w h i c h to design an e x p e r i m e n t b a s e d upon a predicted cephalic r e s p o n s e to a beverage. In previous similar studies, the b e v e r a g e w h i c h ostensibly elicited increased subjective h u n g e r w a s c o n s u m e d in two m i n u t e s (21). H o w e v e r , elicitation o f a cephalic r e s p o n s e d e p e n d s in part u p o n the duration o f

TABLE 6 ENERGY CONSUMED (kcal) BY RESTRAINED (R) AND NONRESTRAINED (NR) EATERS Energy by Macronutrient Treatment A B C D

Group R NR R NR R NR R NR

CHO 608 a 460 d 621 a 436 d 611 ~ 490 d 549 a 467 d

± ± ± ± ± ± -----

PRO 71 31 91 38 69 49 93 43

365 b 260 ¢ 291 b~ 285 ¢ 378 b 287 e 368 b 277 ¢

± 34 --+ 19 ± 51 --+ 31 ± 39 --- 31 +__ 54 --- 26

FAT 982 c 542 f 837 c 621 f 879 c 629 e 810 c 682 f

_ ± ± ± ± ± ± -

Total Energy* 133 45 131 78 77 88 60 77

1970g 1290h 17708 1370h 1880g 1432 h 1740g 1456 h

--- 210 __. 75 ± 231 --- 134 ± 150 ± 148 ± 149 ± 113

*Values shown are m e a n _ 1 SEM. Within a column or within a row, values with different superscripts are significantly different, at p<0.05.

SOFT DRINKS WITH ASPARTAME

the sensory stimulation (6), and soft drinks of the type used in the present study are not generally consumed in two minutes or less. Therefore, it was important to determine whether duration of stimulation, in addition to amount of aspartame consumed, could affect appetite. Thus, not only was the volume of soft drink (and so amount of aspartame) varied, but time allowed for consumption was also varied. This design allowed examination of effects on appetite and feeding behaviour due to both duration of sensory stimulation and the amount of aspartame ingested. However, amount of aspartame ingested was not independent of volume of beverage in this experiment. Since there was no appetite enhancement detected by the VAS measures, as might have been predicted from earlier reports of the effect of aspartame-sweetened water (4,21), it might be argued that the VAS was not completed in a quantitative manner consistent with the subject's feelings of hunger and satiety. This is unlikely for two reasons. First, meals consistently decreased desire to eat, hunger and prospective consumption, and increased fullness. As well, desire to eat, hunger and prospective consumption reached a maximum just prior to lunch, and fullness fell to its minimum value. Second, there was a significant reduction in subjective measures of appetite (and elevation of subjective fullness) detected for the first 30 minutes following consumption of the 560-ml diet soft drink preload compared with the 280-ml diet soft drink or carbonated water preload, indicating that treatment could alter these measures. Furthermore, it should be noted that a recent report by Rogers, Pleming and Blundell (22) failed to repeat the increase in subjective appetite following ingestion of a laboratory-mixed beverage sweetened with aspartame, reported previously (4,21). Rogers et al. (22) suggested that the absence of any appetite enhancement in their recent study was likely due to gulping of the drink by the subjects, reducing the amount of sensory stimulation. However, the present study shows that when subjects slowly consume an aspartame-sweetened beverage over a 10-min period, there is no effect on appetite. In addition, others have allowed subjects anywhere from 5 min (26) to 10 rain (18,26) to consume the beverage offered, and found that aspartame had no unique effects on appetite. The appetite-suppressing effects of consuming 560 ml of a diet soft drink may have been due to the large volume of the beverage, suppressing appetite through stomach distension, or may have resulted from a direct effect of aspartame. Appetitesuppressing effects of ingesting similar amounts of aspartame (234 mg to 470 mg) in pill form, and hence without taste, have been reported (22). That 330 mg of aspartame in a diet soft drink suppresses appetite seems unlikely, however, since we have reported studies which showed that subjects given as little as 800 mg and as much as 10 g of aspartame in pill form failed to demonstrate either an effect on lunch-time intake or on subjective measures of appetite (1,27). Although it is difficult to envisage a physiological mechanism for the food intake suppressive effect of a small, but not a large, amount of aspartame, such direct comparisons have not been made and so the possibility remains to be investigated. The sample size in this study (n = 20) was chosen to allow detection of a treatment effect if it caused a 10% change in lunchtime intake, with a power of 0.80. Our calculations of sample size requirement followed the procedures described by Wirier (34) and were based on similar experiments where an n of 20 was sufficient to detect effects of tryptophan on food intake, as well as the effect of menstrual phase on food intake, assuming an intraindividual variance of 300 kcal (10,11). The negligible effect on feeding produced by consuming a diet soft drink preload, and the relatively large intraindividual variance in amount eaten at each of the lunch-time meals, resulted in no statistically signifi-

809 cant effects of treatment on lunch-time food intake. Using the data derived from this study, which assumes the same weak effect of treatment but a possible statistical significance, the power of the study was only 0.10. Regardless, this low power does not alter the conclusions, nor cause undue concern, for three reasons. First, as indicated above, the study was designed to detect an intake change of 10%; the largest average change observed in this study was just over 5%, the smallest was 1.1%. Furthermore, individual subjects were as likely to decrease their intake as increase it following consumption of the diet pop compared to the mineral water preload. Second, in order to obtain a power of 0.80 with this level of treatment effect, assuming that it was small but consistently reproducible, and with the intake changes and variances observed in this study, over 200 subjects would be required, even in a paired comparison design. If an n of 200 is required to provide absolute assurance that soft drinks have even a small but consistent effect on food intake, and that we have in this study arrived at a false negative, then studies examining the effect of sweetened beverages on food intake cannot be usefully designed based on short-term changes in food intake. Perhaps continued monitoring of food intake over the day would provide some evidence of the hypothesized increase in food intake. Indeed, Rogers and Blundell (20) observed that following consumption of an artificially sweetened yoghurt preload total dally food intake was increased relative to consumption of an unsweetened yoghurt preload. In contrast, however, Tordoff and Alleva (31) found that consumption of an APM-sweetened (caffeinated) soft drink over a 3-week period significantly reduced caloric intake of both males and females (though body weight was significantly reduced in males only). Third, in two of the three comparisons (280 ml of carbonated mineral water vs. 280 ml of diet soft drink, and 280 mi vs. 560 ml of diet soft drink), the mean change in lunch-time intake was in a direction opposite to that predicted if aspartame enhanced food intake, i.e., subjects consumed less energy at lunch following 280 ml of diet pop compared to 280 ml of mineral water, and less following 560 ml of diet pop (largest aspartame dose) compared to 280 ml of diet pop. Therefore the failure to detect an increase in the lunch-time meal intake following a diet soft drink preload compared to a water preload was not due to a lack of power in this study. Rather, it can be stated that aspartame, when delivered in a diet soft drink preload 60 minutes prior to a meal, does not alter feeding behaviour at that meal. Other studies have also failed to detect an intake change following consumption of an artificially sweetened beverage (19, 21, 22, 26). While Rogers et al. (21) argued that this lack of an effect on lunch-time intake was most likely due to a ceiling effect (all subjects being very hungry when served lunch and so eating a maximal amount), a similar argument does not hold for this study for three reasons. First, subjects received a large breakfast (approximately 25% of their normal daily caloric intake) and so were not exceptionally food deprived when lunch was served. Second, subjects consumed the breakfast at 8:00 a.m., a common breakfast time, and the lunch-time meal was served at 12:05 p.m., resulting in a typical intermeal interval. Third, the average size of the lunch-time meal (approx. 1500 kcal) was similar to that observed in other studies (27). Finally, although these results contradict previous reports of increased hunger following consumption of an aspartame-sweetened beverage (4,21) and increased hunger and food intake following consumption of saccharin-sweetened yoghurt (20), they confirm more numerous findings of no direct effect on feeding and appetite following aspartame ingestion (2, 3, 8, 13, 15, 16, 19, 22, 24-27, 29, 31).

810

B L A C K ET AL.

In summary, the consumption o f aspartame-sweetened beverages did not increase short-term subjective hunger, or food intake, in a meal taken within the following 60 to 90 minutes. The

results add support to the growing body o f evidence indicating that the consumption o f foods and beverages containing nonnutrifive sweeteners do not increase hunger and food intake.

ACKNOWLEDGEMENTS This research was supported by the International Life Sciences Institute-Nutrition Foundation, Washington, DC. We would like to thank Cathy Pak for excellent technical assistance.

REFERENCES 1. Anderson, G. H.; Leiter, L. A. Effects of aspartame and phenylalanine on meal-time food intake in humans. Appetite 1l(Suppl.):4853; 1988. 2. Anderson, G. H.; Saravis, S.; Schacher, R.; Leiter, L. A.; Zlotkin, S. Aspartame: effects on appetite and food intake in children. Fed. Proc. 2(5):1197 (No. 5237); 1988. 3. Anderson, G. H.; Saravis, S.; Schacher, R.; Zlotkin, S.; Leiter, L. A. Aspartame: effect on lunch-time food intake, appetite and hedonic response in children. Appetite 13:93-103; 1989. 4. Blundell, J. E.; Hill, A. J. Paradoxical effects of an intense sweetener (aspartame) on appetite. Lancet 1:1092-1093; 1986. 5. Blundell, J. E.; Rogers, P. J.; Hill, A. J. Uncoupling sweetness and calories: methodological aspects of laboratory studies on appetite control. Appetite 1 l(Suppl.):54-61; 1988. 6. Brand, J. G.; Cagan, R. H.; Naim, M. Chemical senses in the release of gastric and pancreatic secretions. Annu. Rev. Nutr. 2:249276; 1982. 7. Dubuc, J.; Lahale, A. Nutritive value of foods. Montreal: Universite de Montreal; 1987. 8. Foltin, R. W.; Fischman, M. W.; Emurian, C. S.; Rachlinski, J. J. Compensation for caloric dilution in humans given unrestricted access to food in a residential laboratory. Appetite 10:13-24; 1988. 9. Herman, C. P.; Polivy, J. Restrained eating. In: Stunkard, A. J., ed. Obesity. Philadelphia: W. B. Saunders and Co.; 1980:208-223. 10. Hrboticky, N. Effects of tryptophan on mealtime food intake of normal weight men and women. Ph.D. thesis, University of Toronto; 1986. 11. Hrboticky, N.; Leiter, L. A.; Anderson, G. H. Effects of L-tryptophan on short term food intake in lean men. Nutr. Res. 5:595-607; 1985. 12. Kanders, B. S.; Kowalchuk, M. B.; Lavin, P. T.; Blackburn, G. L. Do aspartame (APM) sweetened foods and beverages aid in longterm control of body weight? Abstract from 6th Ann. Mtg. Amer. Assoc. Study Obesity, Bethesda, MD, 1989. 13. Kanders, B. S.; Lavin, P. T.; Kowalchuk, M. B.; Greenberg, I.; Blackburn, G. L. An evaluation of the effect of aspartame on weight loss. Appetite 1l(Suppl.):73-84; 1988. 14. Louis-Sylvestre, J.; LeMagnen, J. Palatability and preabsorptive insulin release. Neurosci. Biobehav. Rev. 4(Suppl. 1):43-46; 1980. 15. Mattes, R. Effects of aspartame and sucrose on energy intake and hunger in humans. Physiol. Behav. 47:1037-1044; 1990. 16. Parham, E.; Marshall, S.; Parham, A. Influence of aspartame on hunger ratings and food intake in college students. Abstract from 6th Ann. Mtg. Amer. Assoc. Study Obesity, Bethesda, MD, 1989. 17. Powley, T. L. The ventromedial hypothalamic syndrome, satiety and a cephalic phase hypothesis. Psychol. Rev. 84:89-126; 1977. 18. Powley, T. L.; Berthoud, H-R. Diet and cephalic phase insulin response. Am. J. Clin. Nutr. 42:991-1002; 1985. 19. Rodin, J. Comparative effects of fructose, aspartame, glucose and

20. 21.

22. 23. 24.

25. 26. 27. 28. 29.

30. 31. 32. 33. 34.

water beverages on calorie and macronutrient intake. Am. J. Clin. Nutr. 51:428-435; 1990. Rogers, P. J.; Blundell, J. E. Separating the actions of sweetness and calories: Effects of saccharin and carbohydrates on hunger and food intake in human subjects. Physiol. Behav. 45:1093-1099; 1989. Rogers, P. J.; Carlyle, J.; Hill, A. J.; Blundell, J. E. Uncoupling sweet taste and calories: Comparison of the effects of glucose and three intense sweeteners on hunger and food intake. Physiol. Behav. 43:547-552; 1988. Rogers, P. J.; Pleming, H. C.; Blundell, J. E. Aspartame ingested without tasting inhibits hunger and food intake. Physiol. Behav. 47: 1239-1243; 1990. Rolls, B. J.; Gnizak, N.; Summerfelt, B. A.; Laster, L. J. Food intake in dieters and nondieters after a liquid meal containing mediumchain triglycerides. Am. J. Clin. Nutr. 48:66-71; 1988. Rolls, B. J.; Hetherington, M.; Burley, S. J.; Duijvenvoorde, P. M. Changing hedonic responses to food during and after a meal. In: Kate, M. R.; Brands, J. G., eds. Interaction of the chemical senses with nutrition. New York: Academic Press; 1986:247-268. Rolls, B. J.; Hethetington, M.; Luster, L. J. Comparison of the effects of aspartame and sucrose on food intake. Appetite l l(Suppl.): 62-67; 1988. Rolls, B. J.; Kim, S.; Federoff, I. C. Effects of drinks sweetened with sucrose or aspartame on hunger, thirst and food intake in men. Physiol. Behav. 48:19-26; 1990. Ryan-Harshman, M.; Leiter, L. A.; Anderson, G. H. Phenylalanine and aspartame fall to alter feeding behaviour, mood and arousal in men. Physiol. Behav. 39:247-253; 1987. Stellman, S. D.; Garfinkel, L. Patterns of artificial sweetener use and weight change in an American Cancer Society prospective study. Appetite ll(Suppl.):85-91; 1988. Tordoff, M. G.; Alleva, A. M. Effect on food intake and body weight of drinking soda sweetened with aspartame or high-fructose corn syrup. Abstract from 6th Ann Mtg Amer. Assoc. Study Obesity, Bethesda, MD, 1989. Tordoff, M. G.; Alleva, A. M. Oral stimulation with aspartame increases hunger. Physiol. Behav. 47:555-559; 1990. Tordoff, M. G.; Alleva, A. M. Effect of drinking soda sweetened with aspartame or high-fructose corn syrup on food intake and body weight. Am. J. Clin. Nutr. 51:963-969; 1990. Tordoff, M. G.; Friedman, M. I. Drinking saccharin increases food intake and preference: IV. Cephalic phase and metabolic factors. Appetite 12:37-56; 1989. Van Itallie, T. B.; Yang, M-U.; Porikos, K. P. Use of aspartame to test the "Body weight set point" hypothesis. Appetite l l(Suppl.): 68-72; 1988. Winer, B. J. Statistical principles in experimental design. Toronto: McGraw-Hill Book Co.; 1971.