Metabolic and behavioral effects of a high-sucrose diet during weight loss

Abstract

In response to evidence linking obesity and high amounts of dietary fat, the food industry has developed numerous reduced-fat and nonfat food items. These items frequently derive a relatively large percentage of their energy from sugars and the effect of these sugars on weight regulation is not well known. We studied the comparative effects of high- and low-sucrose, low-fat, hypoenergetic diets on a variety of metabolic and behavioral indexes in a 6-wk weight-loss program. Both diets contained approximately 4606 kJ energy/d with 11% of energy as fat, 19% as protein, and 71% as carbohydrate. The high-sucrose diet contained 43% of the total daily energy intake as sucrose; the low-sucrose diet contained 4% of the total daily energy intake as sucrose. Twenty women aged 40.6 +/- 8.2 y (mean +/- SD) with a body mass index (in kg/m2) of 35.93 +/- 4.8 consumed the high-sucrose diet; 22 women aged 40.3 +/- 7.3 y with a body mass index of 34.93 +/- 4.4 consumed the low-sucrose diet. Mixed-design analysis of variance showed a main effect of time (P < 0.01), with both diet groups showing decreases in weight, blood pressure, resting energy expenditure, percentage body fat, free triiodothyronine (FT3), urinary norepinephrine, and plasma lipids. Small but significant interactions were found between group and time in total cholesterol (P = 0.009) and low-density lipoprotein (LDL) (P = 0.01). Both groups showed decreases in depression, hunger, and negative mood, and increases in vigilance and positive mood with time (P < 0.01). Results showed that a high sucrose content in a hypoenergetic, low-fat diet did not adversely affect weight loss, metabolism, plasma lipids, or emotional affect.

Full text

908 Am J C/in Nutr l997;65:908-15. Printed in USA. tO 1997 American Society for Clinical Nutrition
Metabolic and behavioral effects of a highsucrose diet
during weight 1oss13
Richard S Surwit, Mark N Feinglos, Cynthia C McCaskill, Sara L Clay, Michael A Babyak,
Brenda S Browniow, Claudia S Plaisted, and Pao-Hwa Lin
ABSTRACT In response to evidence linking obesity and high
amounts of dietary fat, the food industry has developed numerous
reduced-fat and nonfat food items. These items frequently derive a
relatively large percentage of their energy from sugars and the
effect of these sugars on weight regulation is not well known. We
studied the comparative effects of high- and low-sucrose, low-fat,
hypoenergetic diets on a variety of metabolic and behavioral
indexes in a 6-wk weight-loss program. Both diets contained
=4606 kJ energy/d with 1 1% of energy as fat, 19% as protein, and
71% as carbohydrate. The high-sucrose diet contained 43% of the
total daily energy intake as sucrose; the low-sucrose diet contained
4% of the total daily energy intake as sucrose. Twenty women
aged 40.6 ± 8.2 y (i ± SD) with a body mass index (in kg/rn2) of
35.93 ± 4.8 consumed the high-sucrose diet; 22 women aged
40.3 ± 7.3 y with a body mass index of 34.93 ± 4.4 consumed the
low-sucrose diet. Mixed-design analysis of variance showed a
main effect of time (P < 0.01), with both diet groups showing
decreases in weight, blood pressure, resting energy expenditure,
percentage body fat, free triiodothyronine (FT), urinary norepi-
nephrine, and plasma lipids. Small but significant interactions
were found between group and time in total cholesterol (P
0.009) and low-density lipoprotein (LDL) (P 0.01 ). Both groups
showed decreases in depression, hunger, and negative mood, and
increases in vigilance and positive mood with time (P <0.01).
Results showed that a high sucrose content in a hypoenergetic,
low-fat diet did not adversely affect weight loss, metabolism,
plasma lipids, or emotional affect. Am J Clin Nutr l997;65:
908- 15.
KEY WORDS Sucrose, weight loss, metabolism, behavior,
affect
INTRODUCTION
Both epidemiologic and experimental evidence link the high
incidence of obesity in the United States with the high fat
content of the American diet (1-6). Consequently, health ad-
vocacy groups, including the American Heart Association and
the American Diabetes Association, have recommended reduc-
ing dietary fat to achieve as well as maintain weight loss (7, 8).
The food industry has responded with a cornucopia of reduced-
fat and nonfat food items, many of which derive a relatively
high percentage of their energy from sucrose or other sug-
ars. However, questions regarding the prolonged effects of
dietary sugars on weight loss or weight maintenance remain
unanswered. Epidemiologic evidence suggests that obesity,
insulin resistance, and non-insulin-dependent diabetes mellitus
(NIDDM) are associated with high-fat diets (1-6, 9), whereas
carbohydrate intake, including sucrose, is negatively correlated
with the incidence of these problems (1, 9). Animal research
has shown that sucrose feeding enhances hepatic lipogenesis,
probably because of the metabolic effects of the component
fructose (10). Some studies in rodents have shown that sucrose
consumption increases adiposity independent of energy intake
(11), whereas others have shown that it does not (12, 13). A
recent review by Hill and Prentice (14) emphasized the need
for experimental data to define the effects of prolonged high
sugar consumption on weight regulation and body fat in
humans.
Another potential problem with low-fat, high-sucrose diets is
that high intakes of carbohydrate, especially sucrose, are pur-
ported to lead to metabolic derangements, particularly with
regard to lipids. A high carbohydrate intake has been shown to
result in elevated plasma triacylglycerol concentrations in some
individuals (15, 16) and to exacerbate hyperglycemia and hy-
perinsulinemia in patients with NIDDM (15). In rodent species,
sucrose or fructose feeding causes an increase in plasma triac-
ylglycerol concentrations (10), and sucrose feeding has been
shown to impair insulin action in comparison with starch
feeding in rats (17). In humans, sucrose has been shown to
decrease plasma concentrations of high-density-lipoprotein
(HDL) cholesterol in normal young men (1 8) and to increase
fasting tnacylglycerol concentrations in NIDDM patients with
hypertriglycendemia (19). Fructose has been shown to increase
plasma concentrations of triacylglycerol and cholesterol (20,
21). Other studies, however, suggest that increased dietary
carbohydrate leads to elevated triacyiglycerol concentrations,
hyperinsulinemia, and hyperglycemia only when fat and en-
ergy contents are not reduced (13, 22). These conflicting re-
IFrom the Department of Psychiatry and Behavioral Sciences, the
Department of Medicine, and the Sarah W Stedman Nutrition Center, Duke
University Medical Center, Durham, NC.
2Supported in part by NIH grant 5 K05 MH00303-l3 from the National
Institute of Mental Health; The Sugar Association, Inc; and the Kellogg
Company, mc, Battle Creek, Ml.
3Address reprint requests to RS Surwit, Department of Psychiatry and
Behavioral Sciences, Box 3842, Duke University Medical Center, Durham,
NC 27710. E-mail: surwi00l@mc.duke.edu.
Received April 8, 1996.
Accepted for publication October 14, 1996.
by guest on July 14, 2011www.ajcn.orgDownloaded from

HIGH SUCROSE INTAKE DURING WEIGHT LOSS 909
group left the study because of hunger (1) and one other
ports indicate that the metabolic result of long-term con-
sumption of high amounts of sucrose (or indeed complex
carbohydrate) on weight loss is not known.
Finally, the effect of sucrose on behavior has been disputed
for many years. Sucrose consumption has long been blamed for
hyperactivity and other alterations in behavior (23-26). Most
studies, however, have focused on the short-term effects of
sucrose. In one long-term study, Wolraich et al (27) found no
differences in the behavior or cognitive function of children
consuming sucrose over an extended period, compared with the
same children when sucrose had been replaced by aspartame or
saccharin. However, the conditions of that study provided
=22% of daily energy intake from sucrose, considerably less
than that potentially consumed by an individual using the
low-fat food items available in the market today.
Although there is little theoretical rationale to support the
notion that sucrose produces behavioral arousal, there are data
to support a theory making the opposite prediction. Carbohy-
drate consumption, in conjunction with a minimal amount of
protein, has been shown to cause an increase in the ratio of
plasma tryptophan to large neutral amino acids (28), which in
turn is associated with an increase in central tryptophan uptake
and brain serotonin synthesis (29, 30). Furthermore, sucrose
has a greater effect than starch. (28). This carbohydrate-in-
duced change in central serotonin activity would presumably
have a tranquilizing effect as opposed to the exaggerated
arousal and hyperactivity typically attributed to sucrose (31).
To determine both the safety and efficacy of high-sucrose
foods during weight reduction, we studied the comparative
effects of high- and low-sucrose, low-fat, hypoenergetic diets
on a variety of metabolic and behavioral indexes in subjects in
a controlled weight-loss program.
SUBJECTS AND METHODS
Subjects
Sixty women 130-200% of their ideal body weights (32)
were recruited through advertising. Volunteers were excluded
if they took any drug that affected the autonomic nervous
system or metabolism (including nicotine) or any psychotropic
agent; if they had a history of significant cardiopulmonary,
neurologic, gastrointestinal, or endocrinologic illness; or if they
participated in a regular exercise program. Subjects were in-
structed to not change exercise patterns during the diet or the
month preceding the diet. Health was assessed at baseline by
history and physical exam, blood counts, serum chemistry
analyses, and electrocardiogram; these procedures were re-
peated at intervals throughout the diet period.
Eight women were eliminated at baseline for the following
reasons: two for smoking, one because of obesity-related sur-
gery, one for being > 200% of ideal body weight, one because
of thyroid abnormalities, one because of a desire to continue a
structured exercise program, and two because of scheduling
conflicts. After the diet intervention began, eight women were
lost from the high-sucrose group: one because of a death in the
family, one because of a work conflict, three because of a
problem with the food, one because of a problem with the study
structure or food, one because of corticosteroid therapy, and
one for unknown reasons. One woman in the low-sucrose
because of a work conflict. The data reported are for the
remaining 42 women (Table 1).
All procedures were approved by the affiliated medical cen-
ter’s Institutional Review Board.
Study design
The study was a 6-wk weight-loss trial that compared the
efficacy, metabolic effects, and behavioral effects of two hy-
poenergetic diets: a low-fat, high-sucrose diet and a low-fat,
low-sucrose diet. Subjects were paired to control for body mass
index (BMI), age, and menstrual status, and were randomly
assigned (within pairs) to either the high- or low-sucrose diet.
The trial was conducted as a controlled feeding study in which
subjects were provided with all meals and snacks for the 6-wk
period. Subjects also received a list of beverages and season-
ings that could be consumed freely. Weekday dinners were
served in a communal dining room with unrestricted seating at
the Sarah W Stedman Nutrition Center; all other meals were
precooked and packaged as “take-out meals.” All subjects
completed a one-page daily diary to document deviation from
the study diets, hunger, health problems, or concerns. A staff
monitor was present daily during dinner to review the diary,
record blood pressure, answer questions, and give support.
Prediet evaluation included measurement of body composi-
tion, resting energy expenditure (REE), thyroid hormones [thy-
roid-stimulating hormone (TSH), free triiodothyronine (FT),
and free thyroxine (FT)], fasting plasma lipid concentrations
[total cholesterol, high-density-lipoprotein (HDL) cholesterol,
low-density-lipoprotein (LDL) cholesterol, and triacylglycer-
oh, fasting serum glucose, and 24-h urinary norepinephrine and
nitrogen. Subjects also completed questionnaires measuring
psychologic characteristics, and a performance task assessing
attention and vigilance. Measurement of thyroid hormone, Se-
rum glucose, and urinary norepinephrine concentrations and
questionnaires were repeated at the midpoint of the diet inter-
vention. Blood pressure was measured twice weekly; weight
was recorded five times weekly. All measures were repeated
during the final (6th) week of the diet.
TABLE 1
Subject characteristics
High sucrose
(n20)
Low sucrose
(n22)
Whites 12 12
Blacks 8 10
Age (y) 40.6 ±8.2’ 40.3 ± 7.3
Body mass index (kg/m2) 35.93 ±4.8 34.93 ± 4.4
Percentage of total body
weight lost 7.3 (4.3-1 1.4)2 7.7 (3.5-11.3)
Menstrual status
Normal cycles 14 14
Oral contraceptive use 1
Estrogen/progesterone
replacement therapy 3 1
No menstrual cycles3 2 7
‘1 ± SD.
2Mean; range in parentheses.
3Includes women taking continuous estrogen or progesterone.
by guest on July 14, 2011www.ajcn.orgDownloaded from

910 SURWIT ET AL
,RE, retinol equivalents; NE, niacinamide equivalents.
Experimental diets
Two diets, each containing ‘4606 kJ/d, were developed by
using the Minnesota Nutrition Data System (NDS, version 2.5)
and served with a 7-d cyclic menu. As shown in Table 2, both
diets contained similar amounts of most essential nutrients
except for starch, total sugars, and sucrose. To ensure that all
subjects received at least the recommended dietary allowance
(RDA; 33) of all essential nutrients, a women’s One-A-Day
multivitamin (Miles, Inc. Elkhart, IN) was taken as a supple-
ment three times per week. As shown in Table 3, conventional
and fresh foods were used to design the menu for both diet
groups. An effort was made to keep the menus for both diet
groups as similar in variety, taste, and appearance as possible,
eg, both diets contained Rice Krispies (Kellogg Co, Battle
Creek, MI) cereal as breakfast on 1 d. However, one cereal was
double-frosted with sucrose whereas the other was coated
white for appearance and treated with aspartame for sweetness.
Snacks were included on some days; pretzels were provided
with the low-sucrose diet and a sugar-sweetened powdered
drink mix was provided with the high-sucrose diet. Common
recipes such as tuna salad (lunch) or meat and vegetables
(dinner) were used to enhance compliance. Food was procured
by using the same sources and brand names whenever possible
to ensure consistent nutrient quality. Every food item was
weighed to an accuracy within 0.5 g.
Body composition
Total body and regional body composition were measured by
dual-energy X-ray absorptiometry (DXA) (model QDR-2000,
software version 7.10 b; Hologic Inc, Waltham, MA). DXA
estimates fat tissues, lean mass, and bone mineral content.
Resting energy expenditure
REE was measured by indirect calorimetry with a ventilated-
hood system (CPXIMAXD; Medical Graphics, Minneapolis).
Subjects were tested as outpatients after a 12-h overnight fast.
Each subject was placed in a supine position, the ventilated
hood was positioned over her head, and she was instructed to
relax, remain motionless, and breathe normally, but to avoid
sleep. Ventilatory data were monitored continuously for 45-60
mm, but REE was estimated by using only the final 30-mm
segment. Energy expenditure was calculated by using nude
body weight and the Weir formula (34) with adjustment for
protein metabolism based on 24-h urinary nitrogen output. Pre-
and postdiet measurements were scheduled at 8-wk intervals to
control for menstrual variation. Prediet REE was measured 2
wk before the start of the experimental diet; postdiet REE was
measured during the last week (week 6) of the diet interven-
tion. All subjects (except two) underwent a preliminary REE
orientation to practice the procedures and thereby reduce anx-
iety, promote relaxation, and yield more valid data during the
actual measurement.
Questionnaires
Subjects completed baseline measures, including the state
form of the Spielberger State-Trait Anxiety Inventory (Mind
Garden, Palo Alto, CA) (35), the Beck Depression Inventory
(Center for Cognitive Therapy, Philadelphia), and the Positive
TABLE 2
Macronutrient and micronutrient contents of the experimental diets-average of a 7-d menu’
High sucrose Low sucrose
Total with Total with
Nutrients Diet alone supplements Diet alone supplements
Energy (Id) 4552.2 -4840.9 -
Protein (% ofenergy) 18.7 -19.3 -
Fat (% of energy) 10.8 -10.6 -
Carbohydrate (% of energy) 73.3 -70.9 -
Sucrose (% of total carbohydrate) 58.0 -6.0 -
Sucrose(g) 121.2 -11.8 -
Total sugars (g) 165.8 -57.9 -
Starch (g) 30.5 -127.7 -
Total fiber (g) 10.4 -14.9 -
Iron (mg) 15.9 27.47 17.9 29.47
Magnesium (mg) 197.2 -226.3 -
Phosphorus (mg) 702.0 -896.8 -
Zinc (mg) 6.84 13.27 6.98 13.41
Calcium (mg) 409 602 638 831
Copper (mg) 0.6 -0.83 -
Selenium (mg) 77.5 -I 10.0 -
Vitamin A (Lg RE) 1218.5 -1009.7 -
Vitamin C(mg) 65.6 91.3 110.0 135.7
Thiamine (mg) 0.95 1.59 1.57 2.21
Riboflavin (mg) 1.44 2.17 1.44 2.17
Niacin (mg NE) 15.51 24.08 19.2 27.77
Folacin (f.Lg) 286.9 458.3 302.8 474.2
Vitamin B-6 (mg) 1.3 2.16 1.4 2.26
Vitamin B-12 (,.Lg) 3.2 5.77 2.7 5.27
by guest on July 14, 2011www.ajcn.orgDownloaded from

High-sucrose diet Low-sucrose diet
Breakfast
Aspartame-coated Rice Krispies
Skim milk
Orange juice
Toasted bagel
Sugar-free jelly
Lunch
Dinner
‘Kellogg Co. Battle Creek, MI.
2Kraft General Foods, Inc., White Plains, NY.
HIGH SUCROSE INTAKE DURING WEIGHT LOSS 911
TABLE 3
Sample daily menu for the high-sucrose and low-sucrose diets
Breakfast
Double-frosted Rice Krispies’
Bran Buds cereal’
1% milk
Kool-Aid powder’
Lunch
Thin-sliced white bread
Deli-sliced luncheon beef
Mustard
Iceberg lettuce
Gelatin dessert
Marshmallows
Sweet iced tea powder
Dinner
Spicy baked cod, cooked
Spinach (salad)
Green onions
Egg whites
Fat-free Italian dressing
Meringue cookies
(g)
21
7
122
34
25
16.8
5
20
117
21.6
17
70
112
3
33.4
26.9
228
White bread
Deli-sliced luncheon beef
Mustard
Iceberg lettuce
Canned peaches, juice pack
Spicy baked cod, cooked
Spinach (salad)
Green onions
Carrots
Fat-free Italian dressing
White rice
French bread roll
(g)
21
I22
263
55
12
65
12.6
5
20
I 37
46.8
56
3
14
I 3.5
175
34
and Negative Affect Scale (PANAS) (36). The Beck Depres-
sion Inventory and the Spielberger State-Trait Anxiety Inven-
tory are well-validated measures that were used to determine
the participants’ state (anxiety and depression) during the in-
tervention period; higher scores indicate higher levels of de-
pression or anxiety. The PANAS, which provides separate
scores for positive and negative affect states, was repeated
weekly to quantify short-term changes in mood.
The Modified Continuous Performance Task (37) was used
to assess the subjects’ attention and impulsivity. This paper and
pencil activity measures the ability of subjects to concentrate
on a task that requires alertness and accuracy for an allotted
period of time. Scores indicate the percentage of correct re-
sponses. Hunger was rated in the daily diary each evening as
I=not a problem, 3 =moderate problem, or 5=significant
problem.
Laboratory methods
Serum glucose was measured with a Kodak Ektachem 250
analyzer (Johnson & Johnson Clinical Diagnostics, Rochester,
NY). Urinary nitrogen was measured with the Boehringer
Mannheim Hitachi 91 1 apparatus (Indianapolis). Urinary nor-
epinephrine concentrations were measured by using HPLC
with electrochemical detection (Bio-Rad Acclaim; Bio-Rad
Laboratories, Hercules, CA). FF3 was quantified by using a
coated-tube radioimmunoassay with an analog tracer (Becton
Dickinson, Orangeburg, NY). Automated chemiluminescent
immunoassay and microparticle enzyme immunoassay meth-
ods were used for measuring TSH (Access Immunoassay Sys-
tem; Sanofi Diagnostics Pasteur, Chaska, MN and Immulite;
Diagnostic Products Corporation, Los Angeles) and FF4 (Ac-
cess Immunoassay System; Sanofi Diagnostics Pasteur and
Imx; Abbott Diagnostics, Abbott Park, IL). Plasma lipid con-
centrations, including total cholesterol, HDL cholesterol [after
precipitation of the sample by dextran sulfate (molecular
weight: 50 000) and magnesium], and glycerol-blanked triac-
ylglycerol concentrations were determined enzymatically on a
Hitachi 91 1 analyzer. LDL-cholesterol concentration was cal-
culated from total and HDL-cholesterol concentrations.
Statistical analysis
Analyses of treatment effects were conducted by using a
mixed-design analysis of variance (ANOVA), with diet group
(high- compared with low-sucrose diet) as the between-sub-
jects factor and time of measurement (eg, baseline and mid-
and posttreatment) as the within-subjects factor. The test of
principal interest in all analyses was the group-by-time inter-
action, which addresses the question of whether the diet groups
differed in the amount of change over time. The estimated
power for detecting a medium-sized effect (‘ 14% of the
explained variance or a 0.5-SD difference in the treatment
effect) for the time-by-group interaction was 0.70, given a =
0.05 and n=42 (38).
In addition to the traditional Pvalues for ANOVA effects,
we present the effect size, ‘rj2, which is a measure of the total
variability “explained” by the given effect in the model (38).
The i2 value is similar in interpretation to R2 in a multiple-
regression model, and provides a means of assessing the abso-
lute magnitude of the effect, and also the relative importance of
effects within and across ANOVA models. Effect sizes are
bound by 0 and 1, with higher values indicating stronger
effects. Five individuals in the high-sucrose group and four in
the low-sucrose group had missing data for one of the seven
measurement times for the psychologic variables; the missing
data point was replaced by the group mean.
by guest on July 14, 2011www.ajcn.orgDownloaded from

.
V
0
I00
98
96
94
92
90
88
86
84
0 1 2 3 4 5 6
TABLE 4
912 SURWIT ET AL
‘Significant time effect, P<0.001.
21 ±SD.
Time (wk)
FIGURE 1. Changes in body weight during the diet intervention. Group
means and SEs are shown for the high-sucrose ( 0 )and low-sucrose ( #{149})
groups.
RESULTS
Subjects
There were no significant differences between groups in
baseline age, baseline BMI, or percentage of total body weight
lost during the intervention (Table 1).
Weight, REE, percentage total body fat, and percentage
trunk body fat
Change in weight for the high- and low-sucrose groups
across the course of the study is shown in Figure 1. There were
no significant differences between groups in mean weight,
REE, percentage total body fat, or percentage trunk fat (Table
4). The time effect was significant for weight (P < 0.001, r2
=0.88), percentage total body fat (P < 0.001, rj2 0.51),
percentage trunk fat (P <0.001, rj2 0.50), REE (P < 0.001,
2 0.54), and diastolic (P >0.001, i2 0.10) and systolic
(P > 0.001, 2 0.10) blood pressure; all scores decreased
over the duration of the study. All group-by-time interactions
were nonsignificant (Table 4), indicating that the groups did
not differ in the magnitude of this decrease over the duration of
the study, ie, there were no treatment effects. As also shown in
Table 4, the proportion of variance explained by the interaction
term was uniformly small for all variables.
Fasting glucose, TSH, FT3, and FT4
No significant group differences were found for fasting
glucose, urine norepinephrine, TSH, VF, or VF4 (Table 5).
There was a significant time effect for norepinephrine (P <
0.001, 0.15) and VF3 (P <0.001, ij2 0.51), with
concentrations decreasing over time. There was a small but
significant increase over time in Ff4 (P =0.001, ‘rj2 0.13).
No significant group-by-time interactions were detected
(Table 5).
Plasma lipids
Mean concentrations of fasting total cholesterol, LDL cho-
lesterol, HDL cholesterol, and triacylglycerol were not signif-
icantly different between groups (Table 6). The time effect was
significant for all lipid measures: total cholesterol (P <0.001,
2 0.63), HDL cholesterol (P <0.001, ij2 0.73), LDL
cholesterol (P <0.001, ‘q2 0.32), and triacyiglycerol (P
0.04, 2 0.10). The time-by-group effect, however, was
significant for total cholesterol (P = 0.009, ‘r2 0.16) and
LDL cholesterol (P =0.014, j2 0.15), with the low-sucrose
group exhibiting a larger decrease than the high-sucrose group
for both of these measures (Table 6).
Psychologic and behavioral variables
There were no significant group differences in mean levels
of hunger, negative affect, positive affect, depression, or anx-
iety, or in the vigilance task (Table 7). The time effect was
significant for negative affect (P < 0.001, tj2 0.47), depres-
sion (P <0.001, q2 0.29), positive affect (P <0.001, 
0.43), and the vigilance task (P =0.005, q2 0.13), with all
subjects improving on these measures. The time effect was also
significant for hunger (P =0.008, ij2 0.08); all subjects
reported lower levels of hunger at the end of the study than at
the beginning. No significant time-by-group interactions were
detected.
DISCUSSION
Three major conclusions can be drawn from this study. First,
a high intake of sucrose from a low-fat, hypoenergetic diet did
not adversely affect weight loss or other metabolic indexes
when compared with an isoenergetic diet in which sucrose was
replaced by starches and aspartame. Both diet groups showed
equal significant reductions in weight, percentage body fat,
Weight, body fat, resting energy expenditure (REE), and blood pressure (BP) at baseline and posttreatment
Measure
Baseline Posttreatment Interaction
term (P)
Interaction
effect size
(i2)High sucrose Low sucrose High sucrose Low sucrose
Weight (kg)’ 96.69 ± 12.622 96.10 ± 13.68 89.74 ±12.51 88.73 ± 13.20 0.64 0.02
REE (kJ/d)’ 6901.09 ± I 104.24 6795.03 ± 1 129.01 5973.92 ± 787.18 5998.27 ± 720.82 0.62 <0.01
Percentage total body fat (%)‘ 49.71 ± 3.52 48.67 ± 3.01 48.54 ± 3.68 47.07 ± 3.81 0.38 0.02
Percentage trunk fat (%)‘ 48.84 ±4.94 47.51 ±5.40 45.93 ±5.84 45.26 ±6.44 0.66 <0.01
Systolic BP (mm Hg)’ 139.5 ±16.02 131.82 ±13.52 127.95 ± 14.59 129.67 ± 11.28 0.99 <0.04
Diastolic BP (mm Hg)’ 74.85 ± 1 1.08 72.82 ± 9.02 71.5 ± 1 1.95 69.10 ± 8.29 0.14 <0.01
by guest on July 14, 2011www.ajcn.orgDownloaded from

HIGH SUCROSE INTAKE DURING WEIGHT LOSS 913
TABLES
Urinary norepinephrine, fasting serum glucose, thyroid-stimulating hormone (TSH), free triiodothyronine (FT), and free thyroxine (‘F,) at baseline and
posttreatment
Measure
Baseline Posttreatment Interaction
term (P)
Interaction
effect size
(tj2)
High sucrose Low sucrose High sucrose Low sucrose
Norepinephrine (mg/24 h)’ 36.61 ± 15.372 37.83 ± 8.97 30.60 ± 10.96 26.05 ± 8.83 0.34 0.03
Fasting glucose (mmollL) 4.97 ±0.70 4.92 ± 0.58 4.87 ± 0.31 4.65 ± 0.03 0.54 0.02
TSH (U/L) 1.44 ± 0.76 1.55 ± 0.74 1.43 ±0.97 1.34 ± 0.59 0.51 0.02
Fr3 (pmol/L)’ 0.06 ± 0.01 0.06 ± 0.01 0.05 ± 0.01 0.05 ± 0.01 0.21 0.04
Fr4 (pmol/L)’ 14.7 ±2.45 15.44 ± 3.22 16.47 ±2.57 17.25 ± 1.93 0.38 0.02
‘Significant time effect, P<0.01.
21 ±SD.
REE, urinary norepinephrine, and FT, as well as an equal
increase in FT4, suggesting that the metabolic effects of these
diets were similar. As noted earlier, there was sufficient statis-
tical power to detect treatment effects of at least a moderate
size, ie, 14% of the variance could be explained-serum lipid
results confirm this. The effect sizes associated with treatment
differences for the remainder of the outcome variables ex-
plained only 1-4% of the variance in the models, suggesting
that the treatment differences were unlikely to be clinically
meaningful.
The second major finding of this study was that neither
high-carbohydrate, low-fat diet had an adverse effect on
plasma lipids. Although subjects in the low-sucrose group had
a somewhat greater reduction in total and LDL cholesterol, the
main effect of the total reduction in both LDL and total
cholesterol for all subjects was much greater than the group
differences. The interaction effect sizes for LDL and total
cholesterol were small (‘q2 0.16 and 0.15, respectively), and
the low-sucrose group had higher baseline LDL and total
cholesterol concentrations. Thus, the meaning of this difference
is questionable. Although we did not assess glucose tolerance
in our subject population, the observed decrease in fasting
glucose argues against any carbohydrate-induced deterioration
in insulin sensitivity in either diet regimen. Women in both
dietary conditions showed similar significant reductions in
systolic and diastolic blood pressure. Interestingly, one 53-y-
old subject who displayed the typical pattern of markedly
elevated plasma triacylglycerol and hyperglycemia at the be-
ginning of the study showed significant decreases in fasting
glucose, total cholesterol, LDL-cholesterol, and triacylglycerol
concentrations, and an increase in HDL cholesterol during the
6 wk of the high-sucrose diet. These findings are consistent
with epidemiologic data (1, 9, 22, 39) and animal data from our
laboratory (13) that clearly show that high sucrose or complex
carbohydrate consumption does not cause obesity, hyperglyce-
mia, or insulin resistance in the absence of dietary fat. Al-
though it is quite possible that sucrose or complex carbohydrate
consumption may produce different effects when total energy
intake is greater, the use of sucrose or other carbohydrate in a
low-fat, weight-reduction program appears both safe and
effective.
Finally, our study shows conclusively that there were no
behavioral sequelae accompanying high intakes of sucrose.
Subjects consuming 43% of their total energy intake as sucrose
showed no differences in any behavioral index measured corn-
pared with those consuming <5% of their energy intake as
sucrose. All subjects also showed significant improvement
over time in indexes of mood and attention span; hunger
decreased equally in both groups and anxiety ratings did not
change significantly. These results confirm numerous other
studies that have failed to find any behavioral effects of sucrose
consumption in humans or animals (40-42). They extend re-
cent findings by evaluating chronic consumption at about twice
the dose of sucrose (percentage of total daily energy intake)
used by Wolraich et al (27). It is therefore clear that sucrose
ingestion does not cause attentional deficit, mood disturbance,
or appetite problems.
In summary, our study failed to find any adverse metabolic
or behavioral effects of high sucrose consumption in a low-fat,
weight-loss diet. The total daily energy intake of our subjects
included a greater proportion of sucrose than would result from
diets incorporating the low-fat, high-sucrose foods on the mar-
ket today. We therefore conclude that the use of sucrose in a
TABLE 6
Plasma lipids at baseline and posttreatment
Measure
Baseline Posttreatment Interaction
term (P)
Interaction
effect size
(tj2)High sucrose Low sucrose High sucrose Low sucrose
Total cholesterol (mmol/L)’ 4.63 ± 0772 4.92 ± 0.84 4.14 ± 0.75 3.94 ± 0.62 0.009 0.16
LDL (mmollL)’ 2.70 ± 0.50 3.04 ± 0.74 2.60 ± 0.62 2.38 ± 0.55 0.01 0.15
HDL (mmol/L)’ 1.35 ± 0.34 1.29 ± 0.22 1.06 ± 0.19 1.03 ± 0.19 0.68 <0.01
Total triacylglycerol (mmol/L)3 1.19 ±0.94 1.29 ±0.71 1.08 ± 0.59 1.05 ± 0.45 0.60 <0.01
‘Significant time effect, P<0.001.
2g ±SD.
3Significant time effect, P=0.04.
by guest on July 14, 2011www.ajcn.orgDownloaded from

914 SURWIT ET AL
TABLE 7
Psychological and be havioral measures t baselin e and posttreatment
Measure
Baseline Posttreatment Interaction
term (P)
Interaction
effect size
(p2)High sucrose Low sucrose High sucrose Low sucrose
Hunger’ 1.93 ± 0.892 1.58 ± 0.63 1.62 ± 0.85 1.26 ± 0.61 0.53 0.02
Vigilance task’ 32.00 ± 7.60 28.52 ± 5.78 33.15 ± 6.71 32.33 ± 4.90 0.19 0.04
Positive affect’ 29.40 ±4.47 29.40 ±6.01 37.09 ± 9.01 36.44 ± 9.05 0.56 0.02
Negative affect’ 21.65 ± 7.55 22.32 ± 12.53 11.71 ± 3.20 14.51 ± 9.44 0.46 0.02
Depression’ 9.05 ± 10.19 6.23 ± 4.00 3.58 ± 4.90 3.57 ± 4.90 0.30 0.03
Anxiety 31.95 ±12.53 30.27 ±9.87 33.82 ± 10.55 29.54 ± 11.52 0.34 0.03
‘Significant time effect, P<0.01.
2g ±SD.
weight-loss regimen is unlikely to cause problems for the
average patient, as long as total energy intake is restricted. fl
We acknowledge the efforts of Kimberly P Hoben, the Sarah W
Stedman Nutrition Center, Duke University, for her assistance in the
development of the experimental diets.
REFERENCES
1. Maron DJ, Fair JM, Haskell WL. Saturated fat intake and insulin
resistance in men with coronary artery disease. Circulation
1991 ;84:2020-7.
2. Lovejoy J, DiGirolamo M. Habitual dietary intake and insulin sensi-
tivity in lean and obese adults. Am J Clin Nutr l992;55:1l74-9.
3. Miller WC, Niederpruem MG, Wallace JP, Lindeman AK. Dietary fat,
sugar, and fiber predict body fat content. J Am Diet Assoc
1994:94:1366-7.
4. Haus G, Hoerr SL, Mavis B, Robison J. Key modifiable factors in
weight maintenance: fat intake, exercise, and weight cycling. J Am
Diet Assoc 1994:94:409-13.
5. Reimer L. Role of dietary fat in obesity. I Ha Med Assoc
l992;79:382-4.
6. Heitmann BL, Lissner L, S#{248}rensenTIA, Bengtsson C. Dietary fat
intake and weight gain in women genetically predisposed for obesity.
Am I Clin Nutr 1995:61:1213-7.
7. AHA Nutrition Committee. American Heart Association guidelines for
weight management programs for healthy adults. Heart Dis Stroke
1994:3:221-8.
8. Powers MA, ed. Nutrition guide for professionals: diabetes education
and meal planning. Alexandria, VA: American Diabetes Association;
Chicago: American Dietetic Association, 1988.
9. Bolton-Smith C, Woodward M. Dietary composition and fat to sugar
ratios in relation to obesity. mt J Obes Relat Metab Disord
1994; 18:820-8.
10. Shafrir E. Metabolism of disaccharides and monosaccharides with
emphasis on sucrose and fructose and their lipogenic potential. In:
Gracey M, Kretchmer N, Rossi E, eds. Sugars in nutrition. Nestle
Nutrition Workshop Series. Vol 25. New York: Vevey/Raven Press,
1991: 13 1-52.
I 1.Keno Y, Matsuzawa Y, Tokunaga K, et al. High sucrose diet increases
visceral fat accumulation in VMH-lesioned obese rats. Int I Obes
1991:15:205-11.
I 2. Esteve M, Rafecas I, Remesar X, Alemany M. Dietary sucrose sup-
plementation fails to modify fat deposition in lean or obese rats. Arch
Int Physiol Biochem Biophys 1992:100:137-42.
13. Surwit RS, Feinglos MN, Rodin I, et al. Differential effects of fat and
sucrose on development of obesity and diabetes in C57BU6i and AJJ
mice. Metabolism 1995;44:645-5 1.
14. Hill JO, Prentice AM. Sugar and body weight regulation. Am I Clin
Nutr l995;62(suppl):264S-74S.
15. Garg A, Bantle JP, Henry RR, et al. Effects of varying carbohydrate
content of diet in patients with non-insulin-dependent diabetes mdli-
tus. JAMA 1994;271:l42l-8.
16. Cole TG, Bowen PE, Schmeisser D, et al. Differential reduction of
plasma cholesterol by the American Heart Association Phase 3 Diet in
moderately hypercholesterolemic, premenopausal women with differ-
ent body mass indexes. Am J Clin Nutr 1992;55:385-94.
17. Storlien LH, Kraegen EW, Jenkins AB, Chisholm Di. Effects of
sucrose vs starch diets on in vivo insulin action, thermogenesis, and
obesity in rats. Am J Clin Nutr 1988:47:420-7.
18. Yudkin J, Eisa 0, Kang 55, Meraji 5, Bruckdorfer KR. Dietary
sucrose affects plasma HDL concentration in young men. Ann Nutr
Metab l986;30:26l-6.
19. Emanuele MA, Abraira C, Jellish WS, DeBartolo M. A crossover trial
of high and low sucrose-carbohydrate diets in type II diabetics with
hypertriglyceridemia. J Am CoIl Nutr 1986;5:429-37.
20. Swanson JE, Lame DC, Thomas W, Bantle JP. Metabolic effects of
dietary fructose in healthy subjects. Am J Clin Nutr 1992:55:851-6.
21. Hollenbeck CB. Dietary fructose effects on lipoprotein metabolism and
risk for coronary artery disease. Am J Clin Nutr l993;58(suppl):800S-9S.
22. Lichtenstein AH, Ausman LM, Carrasco W, Jenner JL, Ordovas JM,
Schaefer El. Short-term consumption of a low-fat diet beneficially
affects plasma lipid concentrations only when accompanied by weight
loss. Arterioscler Thromb l994;l4:175l-60.
23. Prinz Ri, Robert WA, Hantman E. Dietary correlates of hyperactive
behavior in children. J Consult Clin Psychol l980;48:760-9.
24. Crook WG. Food allergy-the great masquerader. Pediatr Clin North
Am 1975;22:227-38.
25. Goldman JA, Lerman RH, Contois JH, Udall JN Jr. Behavioral effects
of sucrose on preschool children. J Abnorm Child Psychol
1986; 14:565-77.
26. Wender E. Review of research on the relationship of nutritive sweet-
eners and behavior. In: Diet and behavior. Washington, DC: National
Center for Nutrition and Dietetics, 1991:65-80.
27. Wolraich ML, Lindgren SD, Stumbo P1, Stegink LD, Applebaum MI,
Kiritsky MC. Effects of diets high in sucrose or aspartame on the
behavior and cognitive performance of children. N Engl J Med
1994;330:301-26.
28. Lyons PM, Truswell AS. Serotonin precursor influenced by type of
carbohydrate meal in healthy adults. Am I Clin Nutr 1988;47:433-9.
29. Fernstrom JD, Wurtman RJ. Brain serotonin content: increase follow-
ing ingestion of carbohydrate diet. Science 197 l;174:1023-5.
30. Femstrom ID, Wurtman RI. Brain serotonin content: physiological
regulation by plasma neutral amino acids. Science l972;l78:414-6.
31. Lieberman HR. Wurtman JJ, Chew B. Changes in mood after carbo-
by guest on July 14, 2011www.ajcn.orgDownloaded from

HIGH SUCROSE INTAKE DURING WEIGHT LOSS 915
hydrate consumption among obese individuals. Am I Clin Nutr
I986:44:772-8.
32. Metropolitan Life Insurance Co. Ideal weight (kg) for height, adults
(modified from 1983 Metropolitan height and weight tables). In:
Nutritional care manual. Durham, NC: Duke University Medical Cen-
ter Department of Dietary Services, 1991.
33. National Research Council. Recommended dietary allowances. 10th
ed. Washington, DC: National Academy Press, 1989.
34. Weir JBV. New methods for calculating metabolic rate with special
reference to protein metabolism. J Physiol 1949:109:1-9.
35. Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA.
Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Con-
suIting Psychologists Press, Inc. 1983.
36. Watson D, Clark LA. Development and validation of a brief measure
of positive and negative affect: the PANAS scales. I Pers Soc Psychol
I988:54:1063-70.
37. Rosvold HE, Mirsky A, Sarason I, Bansome ED, Beck LH. A contin-
uous performance test for brain damage. J Consult Psychol
l956;20:343-50.
38. Judd CM, McClelland GH. Data analysis: a model-comparison ap-
proach. New York: Harcourt Brace Jovanovich, 1989.
39. Ullmann D, Connor WE, Hatcher LF, Connor SL, Flavell DP. Will a
high-carbohydrate, low-fat diet lower plasma lipids and lipoproteins
without producing hypertriglyceridemia? Arterioscler Thromb
1991;l 1:1059-67.
40. Wolraich ML, Wilson DB, White JW. The effect of sugar on behavior
or cognition in children. JAMA 1995;274:1617-21.
41. White JW, Wolraich M. Effect of sugar on behavior and mental
performance. Am I Clin Nutr 1995;62(suppl):242S-9S.
42. Brownlow BS, Petro A, Feinglos MN, Surwit RS. The role of motor
activity in diet-induced obesity in C57BLI6J mice. Physiol Behav
l996;60:37-41.
by guest on July 14, 2011www.ajcn.orgDownloaded from