Separating the Actions of Sweetness and Calories: Effects ...

Phystology&Behavtor,Vol 45, pp 1093-1099 ~ Maxwell Pergamon Macrmllanplc, 1989 Plantedin the U S A 0031-9384/89 $3 00 + 130

Separating the Actions of
Sweetness and Calories:
Effects of Saccharin and
Carbohydrates on Hunger and
Food Intake in Human Subjects I

P E T E R J. R O G E R S 2 A N D J O H N E. B L U N D E L L
Biopsychology Group, Psychology Department, University of Leeds, Leeds LS2 9JT, UK

Received 19 January 1989

ROGERS, P. J. AND J E. BLUNDELL Separatmg the actions of sweetness and calortes. Effects of saccharin and carbohydrates
on hunger andfood mtake in human subjects PHYSIOL BEHAV 45(6) 1093-1099, 1989 - - A comparison was made of the effects
on hunger and food intake of consuming preloads varying m sweetness and energy content The preloads were a plato (unsweetened)
yogurt, and the same yogurt sweetened to equal intensity with sacchann or glucose, or supplemented with starch. This balanced design
made ~t possible to assess the consequences of adding sweetness to food as well as the consequences of substituting a nonnutnt~ve
sweetener for a caloric sweetener Subjects (N = 24, repeated measures design) ate the preload at midday and returned one hour later
for a sandwich lunch Food mtake m this meal was measured directly, and retake during the remaining part of the day was momtored
using home recording m drones. Hunger was assessed usmg subjective ratings of motwatlon to eat Food retake at lunchtlme was
slgmficantly greater following the saccharin compared with the plato preload, and parallel effects were revealed by the motwanonal
ratings Saccharin also stimulated further increases m retake after lunch Food retake was lowest following the high-energy preloads,
w~th the starch supplemented yogurt producing somewhat the largest suppression of retake The results confirm and extend prewous
findings showing that intense sweeteners do not possess the same satiating capacity as glucose and sucrose The snmulatlon of appetite
by sacchann may be due to ~ts sweet taste and also to effects on postmgest~ve mechamsms.

Sweetness Caloric compensation Sacchann Glucose Carbohydrate Food retake Hunger
Appetite Human eating behavlour

INTENSE sweeteners, such as saccharin, cyclamate, aspartame distinguish between the logical consequences of these manipula-
and acesulfame-K, provide a means of sweetening food without tions [see (5)].
adding significantly to its energy value. They can be used as
experimental tools in the study of appeUte control, and in a wider A majority of the experimental stu&es which have examined
context they are perceived by consumers as aids to dietary control. the effects of covertly substituting an intense sweetener for sucrose
Therefore understanding the influences of intense sweeteners on m foods and beverages found that voluntary food intake increased
human appetite has both theoretical and practical significance. to compensate, at least partially, for the lower energy value of the
Two issues which require clarification concern the role of sweet- item(s) containing intense sweeteners [reviewed m (31)]. This
ness and the role o f caloric content. The addition o f an intense demonstrates that ingested calories suppress appetite. However,
sweetener to a food increases its sweetness without altering its nothing can be concluded about the role of sweetness from this
energy value. On the other hand, the substitution of an intense comparison, because sweetness was not systematically varied.
sweetener for sugar in a food reduces the food's energy value Results from stu&es which included a comparison between dif-
while maintaining its sweetness. Many studies have failed to ferent sweetness levels (same calonc value) indicate a stlmulatory
effect of sweetness on appetite. For example, compared with the

Iprehrmnary reports of this study were presented at a meeting of the North Amencan Association for the Study of Obesity, Boston, MA, October, 1987
[(5), includes part of the data from the first rephcatlon only], to the Assocmt~on for the Study of Obesity, London, April 1988, and at a Conference on
the Psychobiology of Eatang Disorders, New York Academy of Sciences, New York, October, 1988.

2p j R is supported by the Agriculture and Food Research Council

1093

1094 ROGERS AND BLUNDELL

effect of water, consumption of intense sweetener solutions was (a) 210 g yogurt, plus 25 g fruit energy value= 131 kcal:
found to promote Increases in subjective hunger, whereas solu- (b) 210 g yogurt, plus 25 g fruit, plus 163 mg saccharin 131
tions of sucrose and glucose of equivalent sweetness reduced
hunger (3, 4, 32) Data reported by Guss, Klsslleff and Pi-Sunyer kcal;
(1 l) show increased food intake following aqueous preloads (c) 170 g yogurt, plus 25 g fruit, plus 50 g starch (maltodex-
containing aspartame (1% glucose and 1% fructose solutions
sweetened with aspartame compared with the same volume of trin, DE = 7-10; 'Snowflake', Cerestar UK Ltd., Manches-
water). Also, short-term Increases in food retake have been ter)'295 kcal;
observed in rats given saccharin to drink (36). Conversely, Brala (d) 170 g yogurt, plus 25 g fruit, plus 50 g glucose (dextrose
and Hagen (6) found that food intake in a test meal was monohydrate). 295 kcal.
significantly reduced when the perceived sweetness of mdkshake Additionally, the second cohort of subjects received a fifth
preloads was suppressed by a Gymnema sylvestre extract (sweet- preload'
ness perception had returned to normal at least 30 minutes before (e) 170 g yogurt, plus 25 g fruit, plus 50 g starch, plus 163 mg
the test meal was eaten) saccharin" 295 kcal.
The yogurt base was a low-fat plain yogurt (Eden Vale, 60
The present study sought to further investigate the role of kcal/100 g), and to each of the preloads was added unsweetened,
sweetness and energy content in the control of appetite by varying canned fruit (33 kcal/100 g) mashed and drained of juice. The
these parameters in a balanced 2 × 2 cell design (5) Hunger purpose of this was to improve the flavour and appearance of the
motivation and food intake were assessed following the consump- preloads Only soft, dark fruits (blackcurrant, blackberry, rasp-
tion of four types of low-fat yogurt snack (preload) unsweetened, berry and strawberry) were used. Subjects received the same fruit,
and the same yogurt sweetened to equal intensity with saccharin or their most preferred, on each occasion The fruit alone did not
glucose, or supplemented with starch. Yogurt provides an excel- appear to slgmficantly sweeten the yogurt. The preloads were
lent medium for these manipulations since it is commonly eaten in mixed 2 to 3 hours before serving and eaten at room temperature.
unsweetened and sweetened forms. Screening ensured that partic- Macronutrlent contents were (a) and (b), carbohydrate = 16 1 g,
ipating subjects liked these types of yogurt equally, so that protein = 12 5 g and fat 2.3 g; (c), (d) and (e), carbohydrate =63 4
differences in sweetness and energy content would not be con- g, protein = 10.1 g, fat = 1 9 g The volume of each of the preloads
founded with differences in palatability was the same 215 ml These formulations were arrived at
following a pilot study which indicated that the palatability (rated
METHOD 'pleasantness') of the five preloads was similar and that preloads
(b), (d) and (e) were of equal sweetness (also see the Results
Subjects section)

Twenty-four undergraduate volunteers, 6 men and 18 women Measures
(aged 18-29 years) took part in the study. They were taking
psychology as a major or subsidiary subject. Two cohorts of 12 The following assessment procedures were used.
subjects attending different classes were run 6 months apart. No 1) Ratings of motivation to eat were made on 100 mm
monetary inducement was offered for their participation, although visual-analogue scales (word anchored at either end) These were:
they were told they would receive a free lunch on each of the (a) "How strong is your desire to eat?" (Very strong-Very
experimental days. All of the subjects were of normal weight for weak);
height (mean and SD Body Mass Index = 20 7- 1.3), and scored (b) "How hungry do you feel"" (As hungry as I have ever
low or moderately low on the Herman restraint scale (mean and felt-Not at all hungry),
SD score = 11.0---4.7). The restraint scale assesses behavioural (c) "How full do you feel"" (Very full-Not at all full),
and attitudinal concern about dieting and short-term weight stabil- (d) "'How much food do you think you could eatS" (A large
ity High scorers were excluded from the study dunng recruitment amount-Nothing at all),
because they are known to display disturbed eating patterns. For This last rating scale is called 'prospective consumption
example, highly restrained individuals have been found, paradox- Subjects were instructed to "mark the line for each question
ically, to increase their intake of food following a preload according to how you feel at this moment, and regard the
identified as high in calories (13). Potentml subjects were also dimensions as the most extreme you have ever felt." Measuring
screened for their food preferences, and only those who expressed from the right-hand end of each scale in mm gave scores ranging
a liking or better (5-point scale ranging from 'dislike extremely' to from 0-100 mm. These scales were developed more than 10 years
'like extremely') for both unsweetened and sweetened yogurts and ago and have since been used in many studies. There is a large
who indicated that they usually ate yogurt at least once a week body of data supporting their validity and reliability [e.g., (30)].
were recruited. 2) Food intake was measured In a 'test' meal eaten at lunch-
time. Subjects were presented individually with a tray contalnmg
Design and Preloads the following prewelghed foods 16 sandwich quarters (wholemeal
bread, spread with margarine and containing 2 cold filhngs chosen
The study took the form of a preload manipulation followed by from turkey, roast beef, roast pork, ham, tuna, salmon, cheddar
the measurement of motivation to eat and food intake. It was cheese, cottage cheese and peanut butter); tomato, cucumber and
conducted according to a within subjects design, and subjects were lettuce, 6 chocolate covered digestive biscuits or 4 small portions
'blind" to its true purpose (the study was presented as "an of apple pie; fresh fruit (one of apple, banana and orange), Still
investigation of the effects of nutrients on taste perception and mineral water was available to drink with the meal. Individual
food choice"). As far as possible individual subjects were tested subjects' menus were chosen according to their preferences stated
on the same day of the week, usually with one week between each at the beginning of the study, and they received the same menu on
treatment. each of the 4 or 5 occasions. The amount of food offered (> 1,500
kcal) was such that the subjects would not normally be expected to
The preloads were administered in a counterbalanced order. All eat it all. They were told that the meal was being provided as a
subjects received the following four preloads: reward for participating In the study and that they should eat to

SACCHARIN AND FOOD INTAKE 1095

TABLE 1 TABLE 2
EXPERIMENTALSCHEDULE
SUMMARYOF ANALYSISOF VARIANCEFOR MOTIVATIONALRATINGS
Procedure ANDFOODINTAKE F RATIOS

Time Prospec- Food
rive Intake
11:50 a.m SubJect amves at the laboratory Source Desire Hunger Fullness (excluding
l155am Ratings of motivation to eat (basehne) of to Consump- preload)a
Variance Eat non
1230pm Preload consumed
12.55 p.m. Sensory and hedomc ratings of preload* Sweetnessb 0 10 0 72 0.45 0 83 5 91t
100pm Ratings of moUvatlon to eat (1 mln) Caloriesb 10 37t 3 60 15 08~ 3 82 26 36:~
SubJect leaves at approx. 12:05 p m. Timec 8 79:~ 8.11t 5 07t 7 97t 151 90:~
130pm S × Cb 0 79 6 74t 0 10 1 41
to bedume Ratings of motlvanon to eat (30 rain) S × T~ 6.56t 4.09~" 1 83 0 84 1.50
Subject returns to the laboratory C x T¢ 0.34 0 42 2 20 0 12 1 27
Ratings of motivation to eat (60 mm) S × C × T¢ 3.91" 3 12 2 36 0.50 0 51
1.19
Test meal
Subject leaves at approx. 1'30 p m aFor food retakes includmg the preloads the effect of calones 1s
Food intake diary nonslgmficant (F = 0.74), the other F rataos are unchanged

*Second cohort of subjects bMotlvatlonal ratings, df= 1,22, food retake, df= 1,20
CMonvanonal ranngs, df= 2,44; food retake, df= 3,60
*p<0 05, 5"p<0.025, :~p<0 001

satisfaction. The meals were eaten in groups of 4 to 6 subjects and of the diary records was carried out blind to the identity of the
in a relaxed setting, but at all times they were closely superwsed. preload conditions. Energy intakes were computed from the diary
records with reference to standard food composition tables and
3. Eating following the test meal was assessed using diary supplements (26, 34, 38), and manufacturers' data. One subject
records. SubJects were provided with type-written sheets contain- (female) failed to complete the motivational ratings as instructed,
ing instrnctlons and spaces for recording the following information: and one female and two male subjects failed to return a full set of
rime at which a meal, snack or drink was consumed, a detailed diary records. Therefore, the results for the motivational ratings
descripnon of the food or drink, and the amount eaten expressed as and food intake are based on data from 23 and 21 subjects,
a weight (from packaging or kitchen scales) or in household respectively. Statistical analysis was carded out using repeated
measures. measures ANOVA, and planned a pnori two sample comparisons
were made using Student's paired t-test (two-tail probabilities).
4. Sensory and affective ratings of the yogurt preloads were
made on 100 mm visual-analogue scales. There were 8 scales. RESULTS
sweetness, sourness, bitterness, thickness, smoothness, sharpness,
pleasantness, and fillingness. Motivational Ratings

Procedure The effects of consuming the various preloads on ratings of
motivation to eat are shown in Fig. 1. ANOVA revealed a
Subjects were instructed to eat their normal breakfast, but not significant effect of calories for desire to eat and prospective
to eat after 9:30 a.m. The timetable of subsequent events is set out consumption, a significant effect of time for all ratings, a
m Table 1. Subjects were asked to eat all of the preload, and this significant sweetness × time interaction for destre to eat and
took between 3 and 7 minutes. Mineral water was available to hunger, and finally a significant 3-factor interaction for desire to
drink with the preload. On leaving the laboratory at 12:05 p.m. eat (Table 2). There were three clear findings. First, the high-
subjects were instructed not to eat or dnnk (except water) again calorie preloads brought about a greater reduction in appetitive
until they returned for lunch. They were also prowded with a set motivational ratings (particularly desire to eat and prospective
of motivation rating scales and told that they should complete them consumption) than the low-calorie preloads. Second, during the 60
at 12:30 p.m. Following the test meal subjects were asked to minutes following the consumption of the preloads and prior to
record on the diary record sheet all food and drink taken up to the lunch there was a recovery in appetitive monvatlonal ratings and a
time they went to bed that mght. decrease in fullness. Third, the recovery in desire to eat and
hunger was greatest following the sweet preloads, with saccharin
Ratings of sweetness, pleasantness, etc. of samples of the producing the largest effect (accounting for the sweetness × time
yogurt preloads were made by the first cohort of subjects one week and sweetness x calories × time mteracnons). Note that the
after the main part of the study had been completed. After tasting initial effect of consuming the saccharin preload was not signifi-
(and swallowing) a small amount (<20 ml) of yogurt the subjects cantly different from any of the other preloads. Subsequently,
immediately completed all 8 ratings before tasting the next however, only the appetive motivational ratings made following
sample. The order of presentation of the different yogurts was the consumption of the saccharin preload increased to above
counterbalanced across subjects. The second cohort of subjects baseline (Fig. 1).
made these ratings during the experimental sessions, immediately
following consumption of the entire yogurt preload. Food Intake

On completion of each half of the study debriefing meetings The results for food intake are shown in Fig. 2. ANOVA of the
were held in which the subjets were fully informed of the purpose cumulative food intakes revealed significant effects of calories,
and nature of the experimental manipulations. The study was
approved by the Leeds University, Department of Psychology
Ethics Committee.

Data Analysis

The initial scoring of the motivation rating scales and analysis

1096 ROGERS AND BLUNDELI_

Desire to eat fabc Prospective consumpt=on
101
........ Da ~ab
'°.•
'• .. .'' -~)b¢ "- a .
"D~Z'' '
b

_c • PLAIN Fullness
o~ H u n a e r # STARCH 20"
• SACCHARIN

• GLUCOSE

"~ 0 = 10"
-10' 0
... ..... • ..,i)a
-20 1 rnm b -10- afte1 rmpmreload
.. -#"" '
after preload ..•c

la

30 60 30 60
m)n mm
mm m)n

FIG 1 The effects of consuming four types of yogurt preload on subjective motivational ratings Statistically
significant (t-test, p<0.05) changes from basehne (1 e , ratings made immediately before the preloads were
eaten) are md~catedby closed symbols Means within the same interval shanng common letters are slgmficantly
different (p<0 05)

sweetness and t~me, but no s~gnlficant mteractxon effects (Table 146 kcal, respectively, largest t(9) = 1.23, p > 0 1 Thus the effects
2). Food intake was lower following the high-calorie preloads of adding saccharin to the starch-supplemented preload were
compared w~th the low-calorie preloads by an amount that fully similar to the effects of adding saccharin to the low-calorie (plain)
compensated for the difference m their energy values This is preload
confirmed by the finding of a nonsignificant effect of calories
when the preloads were included m the calculation of the cumu- Preload Ratmgs
lative food intakes. The sxgnificant effect of sweetness was due to
h~gher food retake following the saccharin preload and, to a lesser Although the procedures differed, no clear differences were
extent, the glucose preload. evident between the sensory and hedonic ratings made by the two
cohorts of subjects, and therefore their data were combined. These
Food retake m the test meal was sigmficantly higher following data were analysed using smgle-factor ANOVA and the results are
the saccharin preload compared with the plato preload, and Fxg. 2. shown in Table 3. It can be seen that the preloads differed
shows that saccharin appeared to promote further increases in food s~gmficantly in sweetness, sourness, b~tterness and thickness
intake during subsequent intervals. Although this was not evident There were no s~gnificant differences m ratings of smoothness,
from the 3-factor ANOVA, a supplementary analysis comparing sharpness, pleasantness and filhngness. There were no differences
only the saccharin and plain preloads (2-factor ANOVA preload on any of the ratings between the two sweetened preloads and also
and time) confirmed a s~gnificant preload × time interaction, between the two unsweetened preloads, except for thickness The
F(3,60) =2 81, p<0.05, in addition to sigmficant mare effects of glucose preload and, to a smaller extent, the starch preload were
preload, F(1,20)=20.23, p<0 001, and time, F(3,60)= 122.9, rated as less thick than the plain and saccharin preloads, which
p<0.001. were rated as s~mdar on this dimension.

Combined Saccharin and Starch Supplement DISCUSSION

Both desire to eat and hunger were sigmficantly higher 60 The present study has revealed several notable findings
minutes following the saccharin plus starch preload compared w~th Supplementing the yogurt preload with carbohydrate suppressed
the starch preload, smallest t(11)=2.40, p<0.05, whereas there motivational ratings and food intake, whereas raising the level of
were no reliable chfferences between the effects of these preloads sweetness appeared to increase appetite. Furthermore, saccharin
on ratings made immediately after their consumption, largest had a particularly pronounced effect, promoting increases in food
t(ll)= 1.41, p>0.1. Food intake in the test meal was higher intake both in the test meal and during subsequent intervals. (Note
following the saccharin plus starch preload than following the that th~s later effect of saccharin cannot s~mply be attributed to
starch preload, although the difference was only marginally possible inaccuracies in the diary method of recording food intake.
sigmficant [mean difference = 176 kcal, t(9)= 2.04, p<0.1] At In a within subjects design any lack of precxslon m this measure-
the end of the last measurement period (1.e., 1 30 a.m ), this ment should not create systematic bias towards one treatment over
difference was 390 kcal, t(9)= 2.05, p<0.1. The corresponding another. Instead it would be expected to contribute no~se to the
differences between the starch and glucose preloads were 134 and

SACCHARIN AND FOOD INTAKE 1097

Preload plus 5.30 pm TABLE 3
lunchtime test meal
MEAN ( - SE) SENSORYAND HEDONIC RATINGSOF THE
FOUR PRELOADS

1000 a (b) 1200 i:i:i:!l Rating Plain Preload p
'.:A:::
.....-~ 1000 Saccha-
.~.::::. Starch nn Glucose F(3,69)
300

600 Sweetness 24_3 a'b 27___4cd 77+_4ax 76+-4b'd 5808 <0001
Sourness 67_+5a'b 58+-5cd 30_5 ax 29_4 bd 16.94 <0.001
4O0 6OO Bitterness 57+5 ~'b 55_+6c 43_+7~ 31_+5bc 4.26 <0.01
¥ Thickness 61+-4~ 51+5 t' 42±4 ~x 737 <0.001
C Smoothness 65 ± 5 60 ± 5 66±5 be 67 - 5 0 66 >0.1
Sharpness 64_+6 58+-5 51 ± 6 2.13 >0.1
o 1.30 am Pleasantness 55+-5 47-+6 62 - 4 56±5 2 14 >0 1
Filhngness 54 ± 4 50 ± 5 54±5 57 ± 5 0 94 >0 1
2 9.30 pm 50±5
55 ± 5
>
Means sharing common superscripts are significantly different (p<0 05,
1900 d 2300 (g)
E t-test)
'= ~!~:~::!~.i:~:i 2100 e f .', N•:..... Note that ratings of the starch plus sacchann preload made by the second
o 17oo ..::p.:
>2::::: .!v:!~..::~.:: cohort of subjects did not differ slgmflcantly from their ratings of the
~19oo saccharin preload

1500 I the energy content of food [reviewed in (14,31)]. The accuracy of
the adjustment varies between studies, although the first panel of
1300 170T0 Fig. 2 demonstrates full compensation m the test meal for the
reduced energy value of the plain and saccharin preloads compared
1" with the starch and glucose preloads. These findings provide firm
evidence for the unconditioned regulation of short-term energy
FIG 2 Food intake at intervals following the consumption of four types of intake. The practical ~mphcatlon is that reduced-calorie foods and
preload unsweetened yogurt (plain), and the same yogurt sweetened to beverages, such as those in which mtense sweeteners are substi-
equal intensity with glucose or saccharin, or supplemented with starch tuted for sugar, will have lower satiating capacity than equivalent
Subjects ate the lunchtlme meal beginning at 1 00 p m , one hour after the high-calorie ztems.
preload was eaten These cumulative retakes include the preloads (indi-
cated by the stippled areas) The standard errors ranged between _+54 The present study was also designed to examine the effects of
(starch) and ±73 (glucose) for the test meal intakes to between ±91 the addmon of sweetness to food. Earlier studies have indicated
(starch) and +- 128 (saccharin) for the intakes at 1 30 a m The superscnpt that sweetness stimulates increases m hunger and food mtake (see
letters indicate the comparisons between plain, starch or glucose versus Introduction), and the results for the saccharin versus plato
saccharin (intakes including the preloads) which were significant or which preloads and the saccharin plus starch versus starch preloads are
approached significance as follows (a) t(20)= 2 42, p<0 05, (b) t(20)= apparently consistent with this conclusion. Food intake was also
1 75, p<0 1, (c) t(20)=3.06, p < 0 0 1 , (d) t(20)=2 30, p<0 05, (e) somewhat higher following the glucose-sweetened preload com-
t(20) = 4 11, p<0 001, (f) t(20) = 2 54, p<0 02, (g) t(20) = 1 84, p < 0 1 pared with the starch preload, but the difference was not statisti-
None of the comparisons between plain, starch and glucose (preloads cally significant.
included) were significant (all ps>0 1)
Several suggestions have been made concerning the mecha-
data, which in turn would mlhtate against the detection of nisms by whzch sweetness could be expected to mfluence appetite
significant treatment differences.) [e.g., (4,8)]. Most relevant to the present findings ts the possible
involvement of preabsorptive digestive responses. These refer to
An important methodologzcal aspect of the study was that the reflexive changes in pancreatic and gastrointestinal secretions
various preloads were equally preferred. Only volunteers who triggered by sensory contact with food [intestinal, as well as
liked both sweetened and unsweetened yogurt were selected for visual, olfactory and oral mediation may be involved (24, 25,
the study. Consequently, the manlpulat]ons of calorie content and 27)]. As well as facdltating the dlsposztion of food by the body, it
sweetness were not confounded wtth differences m palatability. has been proposed that these reflexes, and in particular the
Similarly, the volume of the preloads was held constant. The palatability dependent cephalic-phase lnsuhn response, play an
sensory ratings confirmed that the glucose and saccharin preloads important role in modulating food mgestton (8, 20, 27, 28).
possessed equivalent high levels of sweetness. Supplementauon Therefore, the sweet taste of sacccharin might promote increases
with starch and with glucose (but not with saccharin) also altered in hunger and food intake by stimulating preabsorptlve insulin
the preloads' viscosity. This additional sensory difference, how- secretion. Supporting this interpretation is evidence that saccharin
ever, was probably largely unrelated to the effects on subsequent ingestion stimulates insulin release m humans (12) and rats [(2,
food intake. For example, as far as the cognitive mediation of 28, 37); but see (10)] Furthermore, other studies have found a
satiety is concerned (31), the 'thinner' cons]stency of the starch decline m blood glucose levels following oral saccharin, also in
and glucose preloads might, if anything, be expected to be humans (1, 16, 17) and in rats (7,35). This may zmphcate
associated with lower satiating capacity. increased insulin secretion, although insulin levels were not
measured. [At least one study (9) which has been quoted as not
The results concerning the effect of preload calories confLrms replicating these findings used only 4 subjects and a small 30 mg
the common, although not universal finding, that human subjects dose of saccharin. /'here was, nonetheless, a 5% decline m blood
adjust their voluntary intake in response to covert manipulations of

1098 ROGERS AND BLUNDELL

glucose levels and a 30% increase in lnsuhn secretion over effects on appetite. Nonetheless, m the present study, as well as
baseline]. Jorgensen's study (16) is particularly noteworthy. A short-term effects, sacchann stimulated increases in food retake
reanalysts of his data conducted by the present authors revealed beyond the test meal, many hours after the consumption of the
that a sigmficant decrease (6%) in blood glucose occurred after the preload. This strongly suggests the anvolvement of a postlngestive
ingestxon of 45 mg saccharin &ssolved m 80 ml water, and also action(s) of saccharin It seems unlikely that the taste of saccharin
when the same solution was tasted but not swallowed (11% could have produced such pronounced effects. On the other hand,
decrease). In contrast, neither the ingestion of water nor saccharin the slow rate of absorption of saccharin from the gut is consastent
introduced into the duodenum by tube had detectable effects with a delay between the ingestion of saccharin and the occurrence
Hence, the taste of saccharin appears to be a sufficient stamulus to of postabsorptive effects [Although Jorgensen (16) detected no
produce a fall m blood glucose. Another important aspect of these changes in blood glucose following an intraduodenal saccharin
results is that the maximum changes in blood glucose levels load, measurements were continued for only one hour ]
occurred at between 20 and 60 minutes after saccharin ingestaon
(l, 9, 16, 17) which corresponds closely with the tame course of The possible exastence of postlngestave influences of saccharin
saccharm's stimulatory effect on hunger observed in this and a comphcates the interpretation of the findings with respect to the
previous study (32). role of sweetness. (Note also that the glucose and starch preloads
may differ signaficantly m thear postingestave effects due, for
Tordoff also proposes the involvement of a cephahc-phase example, to differences m their rates of digestion). The extent to
neural reflex underlying saccharin-induced increases in food which orosensory stimulation by saccharin or some other action of
intake observed m rats (36) In short-term tests rats allowed to th~s substance contrthuted to ~ts observed effects on appetite
drink a saccharin solution eat 10-15 percent more food than when cannot be definitely determined from the present study. Nonethe-
given only water to drink. The evidence to date, however, less, a comparison with stu&es on aspartame supports the
supports a role for a cephalic-phase reflex that &rectly influences conclusion that sweetness stimulates appetate, at least in the short
laver metabolism, whale findings indicate that cephalic-phase term As mentioned above, preloads sweetened wath aspartame
insulin secretion in rats drinking saccharin as unrelated to subse- have also been found to promote increases in hunger and food
quent increases in food intake (36) intake (3, 11, 32). Aspartame and sacchann are, however,
chemically unrelated and would therefore be expected to differ in
The above &scussion has focussed on the effects arising from thear postingestive actaons (31). Indeed, one recent study found
tasting saccharin. However, an addition to the stimulation of that a moderate dose of aspartame consumed m capsules s~gnlfi-
orosensory receptors, possible postingestwe actions of sacchann cantly reduced food intake [see (31)] A similar investigatxon of
should be considered. Saccharin maght, for example, trigger the effects of ingesting saccharin without tasting may help to
receptors within the alimentary canal (24,25) or it might act clarify some of the issues &scussed above.
following its slow absorption from the gut (29,33). Although it as
not metabolised, saccharin injected parentally has been shown to In summary, the present results confirm and extend previous
markedly potentiate insulin-induced hypoglycemm (23), and may findings showing that intense sweeteners do not possess the same
influence glucose metabolism through an ability to inhibit the satmtmg capacity as glucose and sucrose Thus the substxtutlon of
activity of glucose-6-phosphatase ](21), but see (22)]. Further- saccharin for glucose failed to reduce overall food intake. Further-
more, a recent study found that low concentrations of saccharin more, addition of saccharin both to the plato and starch-supplemented
increased the release of insulin from isolated pancreatic islets (15), preloads promoted increases m hunger and food intake. The
and a similar result has been reported for the structurally related stimulation of appetite by saccharin may be due to at sweet taste
sweetener acesulfame-K (18,19) As yet there appears to be no and also to effects on postingestwe mechanisms.
direct evadence to link any of these findings with saccharln's

REFERENCES

1 Althausen, T L , Wever, G K Effect of sacchann and of galactose ehotor of preabsorpnve lnsuhn secretion Am J Physlol 246
on the blood sugar Proc Soc Exp Biol Med. 35 517-519, 1937 R88-R95. 1984
11 Guss, J L , Klssdeff, H R . P1-Sunyer, F X The effects of fructose
2 Berthoud, H-R, Berelter, D A , Tnmble, E R , Selgel, E. G , and glucose preloads on food retake Paper presented to the Eastern
Jeanrenaud, B. Cephahc phase, reflex msuhn secreuon neuroanatom- Psychological Association, USA, April 1988
lcal and physiological charactensatlon Dmbetologm 20 393-401, 12 Halter, J , Kulkosky, P , Wood, S , Makous, W , Chen, M , Porte, D
Afferent receptors, taste perception and pancreatic endocnne function
1981 m man. Diabetes 2 414, 1975 (abstract)
3 Blundell, J E , Hill, A J Paradoxical effects of an intense sweetener 13 Herman, C P , Pohvy, J A boundary model for the regulation of
eating In Stunkard, A J , Stellar. E , eds. Eating and ~ts &sorders
(aspartame) on appetite Lancet (x) 1092-1093, 1986 New York Raven Press, 1984 141-156, 1984
4 Blundell, J E , Rogers, P J , Hall, A, J. Artlfloal sweeteners and 14 Hill, A J , Leathwood, P D , Blundell, J E. Some evidence for short
term calonc compensation m normal weight human subjects the
appetite m humans In Bxrch, G G , Lmdley, M A , eds Low- effects of high and low energy meals on hunger, food preference and
calorie products London Elsevier Apphed Science, 1988 147-170 food retake Hum Nutr Appl Nutr. 41A.244-257, 1987
5 Blundell, J E , Rogers, P. J., HIll, A J Uncouphng sweetness and 15 Holmes, M., Marcus, F , Levme, L S The effects of caffeine and
calories methodological aspects of laboratory studies on appeute sodmm sacchann on glucose uptake m a&pose tissue and lnsuhn
control Appetite ll(Suppl ) 54-61, 1988 release from isolated rat pancreatic islets. S Afr J Sc. 84 J33, 1988
6 Brala, P M , Hagen, R L Effects of sweetness perception and 16 Jorgensen, H The influence of sacchann on blood sugar Acta
caloric value of a preload on short term retake Physlol Behav Physaol Scand. 20'33-37, 1950
30 1-9; 1983. 17 Kun, E , Horvarth, I The influence of oral saccharin on blood sugar
7 Deutsch, B Con&tloned hypoglycerma a mechamsm for saccharin- Proc Soc Exp Blol Med 66 175-177; 1947
reduced sensmvlty to msuhn in the rat, J Comp Physlol Psychol 18 Llang, Y , Maaer, V , Stembach, G., Lah~, L , Pfelffer, E F The
effect of artxficml sweetener on lnsuhn secretion. II Stimulation of
86 350-358, 1974 msuhn release from isolated rat xslets by acesulfame K (m vttro
8 Gelselman, P. J , Novm, D The role of carbohydrates m appetate, expenments) Horm Metab Res 19 285-289, 1987

hunger and obesity Appetite 3.203-223, 1982
9 Goldfine, I D., Ryan, W. G , Schwartz, T B. The effect of glucola,

&et cola and water mgesuon on blood glucose and plasma msuhn
Proc Soc. Exp Biol. Med 131 329-330, 1969
10 Grill, H J., Bemdge, K. C , Ganster, D. J Oral glucose is the prime

SACCHARIN AND FOOD INTAKE 1099

19. L~ang, Y.; Stembach, G.; Maler, V , Pfefffer, E F The effect of 30 Rogers, P. J ; Blundell, J E Effect of anorexic drugs on food retake
artlficml sweetener on msuhn secreUon I The effect of acesulfame K and the rmcro-structure of eating m human subjects. Psychopharma-
on msuhn secretion m the rat (studies m vtvo). Horm Metab Res. cology (Berhn) 6 159-165; 1979
19'233-238, 1987.
31. Rogers, P J., Blundell, J E. Evaluation of the influence of intense
20 Loms-Sylvestre, J ; Le Magnen, J. Palatabdlty and preabsorptlve sweeteners on the short-term control of appettte and caloric mtake--a
msuhn release Neuroscl Blobehav Rev 4:(Suppl 1):43--46, 1980 psychobiological approach In Grenby, T H., ed. Progress in
sweeteners. London Elsevier Apphed Science, m press
21. Lygre, D G The lnhlbmon by sacchann and cyclamate of phos-
photransferase and phosphohydrolase actwmes of glucose-6-phos- 32 Rogers, P J., Carlyle, J., Hall, A J , Blundell, J. E. Uncouphng
phatase Blochim. Blophys Acta 341 291-297, 1974 sweet taste and calories Comparison of the effects of glucose and
three intense sweeteners on hunger and food retake Physiol. Behav
22 Lygre, D. G. Inhibition by sacchann of glucose-6-phosphatase 43 547-552, 1988.
effects of alloxan tn wvo and deoxycholate in wtro Can J Blochem.
54 587-590, 1976 33. Sweatman, T W., Renw~ck, A. G The tissue &stnbut~on and
pharmacokmetlcs of saccharin m the rat. Toxtcol Appl Pharmacol.
23 Macallum, B The potentmuon of msuhn by sulphones Can J. 55 18-31, 1980
Res. 26E.232-238, 1948.
34 Tan, S P., Wenlock, R W., Buss, D. H Immigrant foods. Second
24 Me1, N Intestinal chemosensmvtty. Physlol Rev 6.211-237, 1985 supplement to McCance and Wlddowson's the composmon of foods
25. Me1, N , Arlhac, A., Boyer, A Nervous regulation of insulin release London HMSO, 1985.

by the intestinal vagal glucoreceptors J Auton. Nerv Syst 41 35 Thompson, M. M., Mayer, J Hypoglycemlc effects of saccharin m
351-363, 1981. experimental animals Am J Chn. Nutr 7 80-85, 1959
26 Paul, A. A, Southgate, D A. T. McCance and W~ddowson's the
composmon of foods. 4th ed. London. HMSO, 1987 36 Tordoff, M G How do non-nutritive sweeteners increase food
27 Powley, T L. The ventromedml hypothalamlc syndrome, satiety, and retake? Appeme 1l(Suppl) 5-11, 1988
a cephahc phase hypothesis. Psychol Rev 84-89-126; 1977
28 Powley, T. L , Berthoud, H-R Diet and cephahc phase msuhn 37 Valensteln, E S , Weber, M. L. Potentiation of msuhn coma by
responses Am J Chn Nutr. 42991-1002, 1985 sacchann J. Comp. Physlol Psychol 60 443-446, 1965.
29 Renwlck, G. A The dlsposmon of sacchann m animals and man--a
review Food Chem. Toxlcol 33:429-435, 1985 38 Wiles, S. J , Nettleton, P H , Black, A E., Paul, A. A. The nument
composmon of some cooked &shes eaten m Britain a supplementary
food composmon table J. Hum Nutr 34 189-223, 1980