Contemporary Clinical Trials 33 (2012) 247257

Contents lists available at SciVerse ScienceDirect

Contemporary Clinical Trials

journal homepage: www.elsevier.com/locate/conclintrial

Effect of sugar-sweetened beverages on body weight in children: design and

baseline characteristics of the Double-blind, Randomized INtervention study
in Kids
Janne Catharine de Ruyter , Margreet Renate Olthof,
Lothar David Jan Kuijper, Martijn Bernard Katan
Department of Health Sciences, EMGO Institute for Health and Care Research, VU University Amsterdam, Netherlands, VU University Amsterdam, De Boelelaan 1085,
1081 HV Amsterdam, the Netherlands

a r t i c l e

i n f o

Article history:
Received 17 June 2011
Received in revised form 29 September 2011
Accepted 18 October 2011
Trial was deblinded 3 November 2011
Available online 25 October 2011
Keywords:
Randomized double-blind controlled trial
Sugar-sweetened beverages
Children
Obesity
BMI for age z-score

a b s t r a c t
Background: Intake of sugar-sweetened beverages is associated with overweight in observational studies. A possible explanation is that liquid sugars do not satiate and that their intake
is not compensated by reduced caloric intake from other foods. However, evidence from intervention studies for this hypothesis is inconclusive because previous studies were not blinded.
Hence results may have been influenced by expectations and behavioral cues rather than by
physiological mechanisms.
Methods: We designed the Double-blind, Randomized INtervention study in Kids (DRINK) to
examine the effect on body weight of covertly replacing sugar-sweetened by sugar-free beverages. Children were only eligible if they habitually drank sugar-sweetened beverages. We
recruited 642 healthy children (mean age 8.2, range 4.811.9). We designed, tested and produced custom-made beverages containing 10% sugar and sugar-free beverages with the
same sweet taste and look. Children receive one 250 mL can of study beverage daily for
18 months. We perform body measurements at 0, 6, 12 and 18 months. The primary outcome
is the z-score of BMI for age. The maximum predicted difference in this score between groups
is 0.72, which corresponds with a difference in body weight of 2.3 kg.
Discussion: The double-blind design eliminates behavioral factors that affect body weight. If
children gain less body fat when drinking sugar-free than when drinking sugar-sweetened
beverages that would show that liquid sugar indeed bypasses biological satiation mechanisms.
It would also suggest that a reduction in liquid sugars could decrease body fat more effectively
than reduction of other calorie sources.
2011 Elsevier Inc. All rights reserved.

Financial disclosure: Financial support was obtained from The Netherlands

Organization for Health Research and Development (ZonMw) (grant
#120520010; http://www.zonmw.nl/en/), the Netherlands Heart Foundation
(grant #2008B096; http://www.hartstichting.nl/), and the Royal Netherlands
Academy of Arts and Sciences (KNAW http://www.knaw.nl/smartsite.dws?
id=25792&lang=EaNG). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No industry funding is involved.
Corresponding author at: VU University, Faculty of Earth- and Life Sciences,
De Boelelaan 1085, 1081 HV Amsterdam, the Netherlands. Tel.: +31 20
5983521; fax: +31 20 5983668.
E-mail addresses: J.C.de.Ruyter@vu.nl (J.C. de Ruyter), M.R.Olthof@vu.nl
(M.R. Olthof), L.D.J.Kuijper@vu.nl (L.D.J. Kuijper), katan99@falw.vu.nl
(M.B. Katan).
1551-7144/$ see front matter 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.cct.2011.10.007

1. Introduction
Obesity in children has become a major health problem
worldwide. In the past three decades the prevalence of overweight in children has increased dramatically [1]. Recently
the prevalence of high body mass index in children appeared
to plateau [2] but the number of overweight children remains
high. Obesity in children is a risk factor for adult obesity, type
2 diabetes, cancer, cardiovascular diseases and death before
55 years of age [38]. Obesity in children also has negative
health consequences during childhood itself such as insulin

248

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

resistance, hypertension, dyslipidemia and type 2 diabetes

[9].
The increase in prevalence of obesity coincided with a
large increase in consumption of sugar-sweetened beverage
in children [10]. Energy intake among children increased
substantially between 1977 and 2001 in the US, and energy
from sugar-sweetened beverages accounted for more than
50% of this increase [11]. Sugar-sweetened beverages, also
known as liquid sugars, are thought to be more fattening
than sugars in solid form because they do not satiate [12].
As a result children would not compensate for the intake of
liquid calories by eating less of other foods and drinks. Therefore sugar-sweetened beverages might increase total energy
intake and cause overweight.
Indeed, in several observational studies intake of sugarsweetened beverages was associated with weight gain both in
adults [1315] and in children [1618]. However, some studies
failed to find such an association [1921]. Also, an association
between sugar-sweetened beverages and weight gain may be
due to confounding because people who drink more soft drinks
often also eat more fast food and exercise less. Although most
studies adjust statistically for known confounders, residual
and unmeasured confounding are of increasing concern in observational nutrition studies [22,23]. Hence trials are needed
for conclusive evidence about causality.
Results of trials of the effect of liquid sugars on overweight are inconclusive, with some finding an increase in
weight [2427] while other trials found equivocal results
[2830]. The lack of conclusive evidence may be due to a
lack of proper placebo treatments, small samples sizes,
short duration and lack of individual randomization. Also,
none of these trials was properly blinded. Blinding is crucial
because without it, results may be biased by effects on food
intake of behavioral cues and expectations rather than by
physiological mechanisms [31].
Hence there is a need for conclusive evidence whether
liquid sugars fail to evoke compensatory changes in food intake
and are therefore more fattening than other sources of calories
[32,33]. We designed DRINK the Double-blind, Randomized
INtervention study in Kids to examine the effect of covertly
replacing sugar-sweetened beverages by sugar-free beverages
on body weight. This allows us to study biological compensatory mechanisms independent of behavioral cues and voluntary
changes in intake. In this article we describe the rationale,
design, methods and baseline characteristics of this study.

conducted in the Netherlands at eight schools in an urbanized

area near Amsterdam. We distribute the beverages to the
schools (bi)weekly by courier. The study was preceded by a
pilot study to test feasibility and logistics and by a sensory
study (see Formulations) to help us in developing the beverages.
2.2. DRINK pilot study

2. Methods

The pilot was a two-month study conducted in AprilMay

2009 at one elementary school in Purmerend, 20 km from
Amsterdam. The study was conducted to the guidelines laid
down in the Declaration of Helsinki and the study protocol
was approved by the Medical Ethical Committee of VU University Medical Centre Amsterdam. Written informed consent was obtained from a parent or guardian. In the rest of
this paper parent will refer to both parents and guardians.
The inclusion and exclusion criteria for the pilot were the
same as for the main study (see Recruitment). Forty-four
children aged 512 were randomized to receive daily one
330 mL can with either sugar-sweetened or sugar-free beverage (bought from Unilever, Colworth, UK). Both were available with lemon and peach flavor. Our courier transported
the beverages to the school.
We asked various stakeholders for advice on how to set up
the main DRINK study. Frequent conversations with the children taught us that offering more than two flavors would
help to increase compliance, and that newsletters, contests
and birthday cards would be appreciated. Therefore we
implemented these and other forms of entertainment in the
main study. The pilot lacked entertainment except the possibility to collect fancy stickers that we attached to the cans in
the last two weeks. In an evaluation form teachers reported
that it was not a problem for them to accommodate consumption of our beverages at school. The parents reported that the
children liked the pilot and that they were keen on bringing
home the weekend and holiday cans. Parents also brought
up that it would be impossible to transport cans during the
summer holidays of 6 weeks. Therefore we interrupted treatment in the main DRINK study for 6 weeks in the summer of
2010, and prolonged it by 6 weeks at the end. Parents also
did not like to return empty cans used at home back to school
during the pilot. Hence we collect cans used at school but not
cans used at home to assess compliance during the main
study (see Ancillary measurements). Our advisory board
suggested that schools might expect something in return for
their participation. Therefore we offered them health related
education materials. Two schools made use of this option.

2.1. Outline

2.3. Design and setting of main DRINK study

DRINK is an individually randomized, double-blind, controlled, parallel intervention study in free-living school children. For 18 months, 642 school children randomly receive
daily either one can of sugar-free beverage (treatment) or
one can of sugar-sweetened beverage (control). We would
have preferred to use water as the treatment beverage, but
that would make blinding impossible. Children consume the
beverage at school on weekdays during their morning break,
and at home during weekends and holidays. The study is
double-blind; neither the researchers nor the children, parents
or teachers know who drinks which beverage. The study is

The study protocol was approved by the Medical Ethical

Committee mentioned above. DRINK is an individually randomized, double-blind, controlled, parallel intervention
study in free-living school children aged 511 years (Fig. 1).
All children receive 250 mL of the study beverage per day
which provides either 0 g or 25 g of sugar. Children in the
control group do not change their liquid caloric intake because they replace their customary sugar-sweetened beverage brought from home by a sugar-sweetened study
beverage. Children in the treatment group replace their customary sugar-sweetened beverage by a sugar-free study

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

249

Fig. 1. Flow of recruitment and design of the Double-blind Randomized INtervention study in Kids (DRINK) that studies the effect of sugar-sweetened beverages
on body weight.

beverage. Therefore they consume 100 liquid calories per day

less than the control group.
We conduct the study at eight elementary schools, in the
Zaanstreek, Purmerend and Haarlem, in an urbanized area,
1633 km from Amsterdam. It is customary for children in
Dutch elementary schools to consume a beverage brought
from home in class during a morning break around 10 am
under supervision of the teacher. We replace the beverage
brought from home by a study beverage and ask the teacher
to check if children consume their beverage during the morning break and to remind them to take home weekend and
holiday cans. Parents supervise consumption of the study
beverage on weekend days and during holidays.

sucralose (0.157 g/L) and Acesulfame-K (0.057 g/L). These

sweeteners are approved by the US Food and Drug Administration, the Joint Commission of Experts on Food Additives of
the World Health Organization, the Food and Agriculture
Organization, and the European Food Safety Authority. The
acceptable daily intake (ADI) for sucralose is 15 mg/kg bodyweight/day [34]. This means that a child of 20 kg can drink
seven cans of sugar-free study beverage per day. The ADI
for acesulfame-K is 9 mg/kg bodyweight/day [35] which corresponds with 13 cans. Since we offer one can per day, the
amount of artificial sweeteners is safe.
We use vegetable and fruit concentrates as colorings in
the mango, peach and raspberry flavored beverages. The
lemon drinks do not contain colorings.

2.4. Beverages
2.4.1. Formulations
The development and manufacture of the beverages was
the major challenge for our study. We needed pairs of beverages, one sugar-free and one sugar-sweetened, which tasted
and looked the same. The beverages needed to be safe and
stable, attractive and pleasant-tasting for children. They also
had to be acceptable for parents and schools. We opted for
a non-carbonated beverage because of the poor health
image of fizzy drinks. We offered several flavors raspberry,
lemon, peach and mango to increase compliance.
Before the study started we tested the sensory qualities of
these beverages in 89 children aged 512 in Purmerend at
the same school where we did the pilot study. The children
liked the sugar-free beverages and sugar-sweetened beverages equally.
The sugar-sweetened beverages contain 25 g of sugar and
100 kcal per 250-mL can (Table 1). The sugar-free beverages
do not provide calories. They contain the artificial sweeteners

2.4.2. Production
We hired JJM (Brussels, Belgium) to design the cans,
Crown (Wantage, United Kingdom) to make the cans, and
Table 1
Composition of study beverages in the Double-blind, Randomized INtervention study in Kids (main DRINK study) a,b.

Sugar-sweetened beverage
per 250-mL can
Sugar-free beverage
per 250-mL can

Sweetener

Calories

Sugar 25 g

100

Sucralose 0.04 g
Acesulfame-K 0.01 g

a
Ingredients besides the sweetener: fruit flavor (lemon, mango, raspberry or peach), fruit juices (lemon, mango, raspberry or peach), colorings (GNT
Exberry vegetable and fruit concentrates; no color in lemon), water, trisodium citrate, citric acid, ascorbic acid, malic acid (malic acid only in sugarsweetened beverage).
b
We bought the colorings from GNT (Mierlo, The Netherlands) and flavorings from Unilever (Colworth, UK). All other ingredients were from
Refresco Benelux (Maarheeze, Netherlands).

250

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

Refresco Benelux (Maarheeze, Netherlands) to produce the

beverages and fill the cans. Fig. 2 presents the can design.
We used 250 mL cans because that fitted better with the factory's logistics than 330 mL cans and because the shelf life of
cans is better than that of cardboard containers. Each of the
four flavors was produced as a sugar-sweetened and a
sugar-free variant. The first batch consisted of 30 000 cans
per flavor/sweetener combination for a total of 240 000 cans,
plus 20 000 cans of sugar-sweetened peach beverage for the
run in period. These were produced on 24 August 2009,
prior to the start of the study. The second batch of 200 000
cans, i.e. 25 000 per flavor/sweetener combination, was produced on 7 September 2010. The beverages will never be
available commercially. VU University has registered the
brand name (Blikkie, colloquial for little can) and the can
design to prevent commercial exploitation by third parties.
2.5. Sample size calculation
The sample size calculation was based on the expected effect of treatment on the change of the z-score BMI for age and
its SD. BMI stands for Body Mass Index and is the weight divided by height squared (kg/m 2).The z-score is the number of
standard deviations by which a child differs from the mean
BMI of children of the same age and gender. We calculated
the number of children required as N = 7.9 2 (withinperson SD of the change in z-score with time/expected difference of change in z-score) (power= 0.8; = 0.05) [36].
We calculated the expected difference of the change in
weight and z-score between treatments from the difference
of sugar intake. Children in the treatment group reduce their
intake of liquid sugar by 25 g per day which equals 12 600 g
of sugar or 50400 kcal in 18 months. A naive calculation
would suggest that at 7709 kcal per kg weight change [37]
this would equate a difference of 6.54 kg. However, any reduction in weight reduces the basal metabolic rate and the energy
costs of activity [38]. Therefore we predict that the mean difference of change in weight between the treatment and control
group is 2.3 kg, if children receiving sugar-free drinks do not
compensate for the absence of sugar by eating and drinking
more of other foods [39].

We converted the predicted difference of change in

weight into a BMI for age z-score because in children the relation between BMI and body fatness depends on age and
gender, while the BMI for age z-score does not. Hence the
BMI for age z-score or standardized BMI allows comparison
of children of different ages and gender.
We estimated the difference of change in BMI for age z-score
from the expected difference in weight change. At the end of
the study our subjects will on average be aged 9.5 years. The
mean weight of Dutch children aged 9.5 years is 32.25 kg,
their mean height is 1.408 m and mean BMI 16.24 kg/m2 [40].
A weight reduction of 2.3 kg will produce a weight of 29.95 kg
and a BMI of 15.11 kg/m2. Thus the maximum expected effect
of the treatment is a reduction of 1.13 kg/m2. In Dutch children
of this age, a fall of 1.13 kg/m2 in BMI will shift a child downwards in the distribution by 0.72 standard deviations [40].
The effect of removing sugary drinks on health could still be
relevant if only part of the sugar removed was compensated by
increased intake of calories from other sources. Also, we cannot
expect 100% compliance to treatment. We designed the study so
that we would still pick up a difference if children in the sugarfree group compensated 50% of the calories lost from sugar by
increased intake of other calorie sources, and if all participants
only consumed 50% of the drinks. The predicted difference of
change in z-score then becomes 50% 50% 0.72 =0.18.
We calculated the within-person SD of the change in zscore from the changes in height and weight of 1017 participants in the ChecKid study [41]. ChecKid examines trends in
overweight and obesity among children in Zwolle initially
aged 412 years. Using raw data obtained from children
when they were aged 6.7 and 9.7 we calculated that the
change in BMI for age z-score over three years, relative to
the Dutch mean averaged 0.03 and its SD equaled 0.66.
Hence our sample size calculation is: N = 7.9 2 (0.66/
0.18) = 212 subjects per treatment. We recruited 642 because we anticipated a dropout of one third.
2.6. Recruitment
We recruited participants between August and November
2009 (Fig. 1). We first approached the boards of seven school

Fig. 2. Can design of the sugar-free and sugar-sweetened beverages. Blikkie is colloquial Dutch for little can. Telephone number and email address of the research team and logo of the study and VU University are depicted on the right side of the can. Expiration date is imprinted at the bottom of each can.

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

districts that encompassed 131 schools for permission to

contact their schools. We approached another 11 schools directly. Out of these 142 schools, 22 were willing and able to
cooperate and granted us access to parents and children.
The 4913 pupils at these 22 schools received a recruitment
flyer including a consent form. We set up information booths
at the schools to tell parents and children about the study.
We also let them taste the study beverages and we sent out
a press release which led to broad exposure in the media.
An independent physician was available for parents to answer health related queries. Parents of 1435 children (29%)
gave written informed consent for their child to participate.
We asked children to give verbal assent to the parent. Parents
that gave their consent were asked to complete a questionnaire so we could determine if the child was eligible. We considered it unethical to increase the children's sugar intake.
Therefore an important inclusion criterion was that children
had to habitually consume sugar-sweetened beverages during the morning break at school on at least three out of five
school days. Other criteria were:

251

2.7. Baseline characteristics

Most of our 642 participants have a normal weight
(Table 2). Their mean BMI for age z-score is 0.03 where 0 is
the Dutch average in 2009. The number of non-western children in our population is slightly higher than that in the
Dutch population as a whole [44]. In 44% of the households
the highest education level attained by either parent is
Intermediate vocational education or less, and 55% had a
Highschool or college degree. The level of education is slightly higher than the Dutch average [45].
2.8. Randomization of participants and beverages
Randomization was done by a statistician (L.D.J.K.) who is
not involved in the execution of the study. Eligible children
were individually randomized. An Excel visual basic macro
Table 2
Baseline characteristics of the 642 children in the Double-blind Randomized
INtervention study in Kids (DRINK).

Inclusion criteria:
Habitual consumption of 250 mL or more per day of sugarsweetened beverages on at least three out of five school
days
Minimum age of 5 years at the start of DRINK
Young enough to be still in elementary school at the end of
DRINK
Written informed consent by a parent
Exclusion criteria:
Medication or medical treatment for obesity
Diabetes, growth disorders, celiac disease, or serious gastroenterologic diseases, e.g. inflammatory bowel disease
Medical history or surgery known to interfere with the
study
Participation in another intervention study up to 3 months
before and during DRINK if that interfered with our study
Physical disabilities that interfere with the measurements
Plans to relocate and change school during DRINK
Because 1435 children volunteered and we only needed
640, we excluded schools that were difficult to reach or had
few prospective participants. This left 699 participants. Application of the inclusion and exclusion criteria left 642 children
to be enrolled and randomized (Fig. 1).
The questionnaire also contained demographic and ethnic
questions about the child and its parents. A child is considered
Dutch if both parents are born in the Netherlands, Non-western
if one or both parents are born in Suriname, Dutch Antilles,
Turkey or Morocco, or Other if both parents are born in a country other than these. In case of a single parent household, we
used the country of birth of this parent. We determined educational level based on both parents, whichever was highest.
Level of education was graded as: Elementary school, Lower
vocational secondary education or technical secondary education, Intermediate secondary education, Intermediate vocational education, High-school graduate and Higher vocational
education/college degree. We used international cut-offs for
overweight and obesity in children and for low and healthy
BMI [42] [43].

Characteristics

n (%) or Mean SD

No. of children
Girls
Age (years)
Ethnicitya
Dutch
Non-western
Other
Household highest educational level
(Dutch translation)b
Elementary (primary) school
(Lagere school/basisonderwijs)
Lower vocational secondary education
or technical secondary education
(LBO, LTS, LEAO, Lagere tuinbouw)
Intermediate secondary education
(MAVO/MULO)
Intermediate vocational education
(MBO, MBA, LO-akten, MTS, MEAO)
Highschool (HAVO/VWO, MMS, HBS)
Higher vocational education/college
degree (HBO/Universiteit)
Weight (kg)
Height (cm)
BMI
BMI for age z-score
Weight statuse
Low BMI
Healthy BMI
Overweight
Obese
Skinfold thicknessc
Biceps (mm)
Triceps (mm)
Subscapular (mm)
Supra-iliac (mm)
Sum of skinfolds (mm)
Waist-to-height ratio (%)
Electrical-impedance fat massd
(kg)
(%)

642
301 (46.9%)
8.2 1.9

502 (78.2%)
120 (18.7%)
12 (1.9%)

6 (0.9%)
30 (4.7%)

51 (7.9%)
195 (30.3%)
92 (14.3%)
259 (40.5%)
30.19 8.86
132.5 12.6
16.8 2.6
0.03 1.04
5 (0.8%)
514 (80.0%)
101 (15.7%)
22 (3.4%)
6.8 3.2
11.8 5.0
8.2 4.8
9.2 5.8
36.0 17.8
44.4 4.0
5.74 3.75
17.83 6.87

n = 634; eight households refused.

n = 633; nine households refused.
n = 641; one child refused.
d
n = 638; four children refused.
e
We used international cut-offs for overweight and obesity in children
and for low and healthy BMI [42] [43].
b
c

252

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

program randomly assigned children to sugar-sweetened or

sugar-free beverages within each school so that mean age,
gender and initial BMI were equal between treatments. Randomization was done with minimal human intervention, as
follows. Data of children who met all inclusion and exclusion
criteria were closed out in the central database (MS
Access) and exported per school into an MS Excel workbook.
The Excel macro written by L.D.J.K. assigned all children a
randomization ID number. Children in the same household
received the same randomization ID number and thus the
same treatment, because there is a risk that they drink each
other's study beverages at home. Children, parents and
teachers have not been informed that siblings receive the
same treatment.
The macro then assigned to each randomization ID number
a random number sampled from a continuous uniform distribution between 0 and 1, e.g. 0.64451. The macro sorted children by this random number, and the upper half was
assigned to one treatment and the lower half to the other treatment. The program then calculated the following percentages
per treatment group: males; birth year 19981999; birth year
20002001; birth year 20022004; BMI b 15; BMI 1518 and
BMI > 18. Ideally all these percentages should come out 50/
50. If any of these percentages was b45% or >55% the macro
automatically assigned new random numbers and reiterated
the process until gender, birth year and initial BMI were evenly
divided between the treatments, i.e. within each school each
treatment group contained more than 45% and less than 55%
of each gender, birth year category and initial BMI category.
The macro ran from start to finish without human intervention.
It reported the final randomization assignments, the percent
distributions of gender, birth year and BMI over treatments,
and the number of iterations needed to achieve this. The statistician assigned the upper and lower half of the random number
distribution to intervention or control treatment by a one-time
flip of a coin for the first school randomized; this assignment
then held for all schools. When a school had an odd number
of participating households the assignment of the median
child or household alternated from one school to the next.
This procedure is equivalent to blocked randomization [46],
but its automation is more straightforward and transparent.
The statistician sent the treatment codes to the database manager who entered them into the central administration database. The database manager is not further involved in the
study.
L.D.J.K also produced anonymous codes for the beverages.
He randomly assigned the codes 1001 to 1008 to the eight
types of beverages (4 sugar-sweetened, 4 sugar-free) and
sent these to the factory. Beverages for the run-in period (all
sugar sweetened) received code 1009. The factory printed the
beverages codes on the cardboard trays on which cans were
packed. The bottom of each can was imprinted with a 17digit code into which the beverage code was encrypted. This
encryption produced 100 different can codes per beverage
type, for a total of 800 can codes.
Personnel at the warehouse where cans are weekly
repackaged for the children are inevitably aware which 4
codes together form one treatment group, but do not know
what the treatment is. When cans are repackaged they are
separated from their cardboard trays, and the codes on the
bottom of the cans are almost always different between

cans even if treatments are the same. Only our statistician

can track down the treatment code for an individual can. So
for parents, teachers, children and study researchers the
codes on individual cans cannot reveal the treatment.
Treatment assignments and beverage codes are held by
L.D.J.K. and by the outside database manager. The University's
attorney holds a print of treatment assignments and beverage
codes in a sealed envelope.
2.9. Packaging and distribution of study cans
The cans were shipped from the factory to the warehouse
of our courier in Egmond aan den Hoef, 50 km northwest of
Amsterdam, and stored at ambient temperature. Between
April and July 2011 the beverages will be stored at 18 C to
prevent deterioration of taste.
At the warehouse, one researcher coordinates the repackaging of the cans according to a standardized procedure. We
package the cans per week per child into 11 21 15.5 cm
cardboard kid boxes imprinted with the study logo. A kid
box holds 8 cans: 5 for school days, 2 for weekend days and 1
spare can. Each kid box also contains 8 name labels that children stick onto their cans each Monday morning to avoid
mix-ups in class. The kid boxes are delivered to the schools in
35 44 16 cm cardboard class boxes, which hold up to 6
kid boxes. Our courier distributes the boxes bi weekly or
every week if the school lacks storage space and brings back
empty and unused cans. We distributed extra spare cans to
all children at the beginning of the study to be used in case children forget to bring home their weekend and/or holiday cans.
After the second production in September 2010 we distributed
new spare cans.
2.10. Body measurements
Measurements are done according to a standardized protocol at t = 0, 6, 12 and 18 months. Children are measured in
their underwear during school hours, generally between
8.30 am and 3.30 pm. We ask children to visit the bathroom
before the measurements. Children who stopped consuming
the study beverages continue to be measured if the parent
and the child permit us to do so.
Two female researchers are always present simultaneously
at each measurement as chaperones. We measure body weight
to the nearest 0.1 kg on a Marsden MPMS-250 digital scale
(Oxfordshire, United Kingdom). We calibrate the scale each
measurement day with a KERN 366-98 standard weight of
20 kg (Balingen, Germany). We measure height to the nearest
millimeter with a SECA 214 (Hamburg, Germany). Height will
be measured in duplicate at 18 months. We measure waist circumference twice to the nearest 0.1 cm at the midpoint between the bottom rib and the top of the hipbone with a SECA
201 flexible steel tape measure (Hamburg, Germany) [47].
We measure biceps, triceps, subscapular, and supra-iliac skinfolds in triplicate with a Harpender Skinfold Caliper HSK-BI
(Burgess Hill, United Kingdom) [47]. Two specially trained researchers perform the waist circumference and skinfold measurements. Each child is measured by the same researcher
throughout the study to reduce variation. We measure armto-leg electrical impedance twice with a BodyStat 1500MDD
(Douglas, United Kingdom). We put 3M 2330 electrodes (St.

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

Paul, U.S.A.) on the right hand and foot according to the manufacturer's manual. Then we connect the two cable leads of the
BodyStat to the electrodes and perform the measurements
according to standardized procedures [48]. We calibrate the
BodyStat each measurement day with the BodyStat calibrator.
The calibrator is a small device onto which we connect to the
cable leads to perform a measurement. The calibration has succeeded if the impedance is between 496 and 503 .
2.11. Ancillary measurements
We measure compliance by counting returned cans and
by analysis of sucralose in urine. We count cans from school
days only, because parents do not return cans used at home
during weekends and holidays. On school days children
place their empty cans back into their kid boxes, and these
are collected by our courier together with the unused cans.
We check returned cans one week each month i.e. 25% of all
cans returned. Cans are scored as empty, half-filled or full.
We will measure sucralose in urine of the children in the
sugar-free group. A study in humans showed that 14.5% of
ingested sucralose is excreted in urine [49]. Therefore we
can use sucralose in urine as a compliance marker. We will
also measure urine samples from children in the sugarsweetened group as a control. Compliance is assumed to be
similar in both groups since the sugar-sweetened and the
sugar-free beverages were equally appreciated by the children (see Formulations). Therefore we collect spot urine
samples from all children at t = 0, 6, 12 and 18 months.
We administer a short dental questionnaire and a hedonic
questionnaire at t = 12 and t = 18 months. The dental questionnaire inquires about the number of new dental fillings and
teeth newly extracted because of caries. The hedonic questionnaire asks: 1. How much do you want to drink the study drink?
2. How satiated do you feel? 3. What do you eat together with
the study drink? 4. How much do you like the study drink?
2.12. Endpoints
Our null hypothesis is that children in the sugar-free group
will fully compensate for the loss of the sugar from their habitual drinks by increasing their intake of calories from other
foods and beverages. Under this hypothesis we will find no difference between the sugar-free and sugar-sweetened group in
the change in BMI for age z-score between t = 0 and
t = 18 months. The alternative hypothesis is that compensation is not 100% and that children receiving sugar-free beverage will gain less body fat than those in the control group.
The primary endpoint is the difference between the
sugar-free and sugar-sweetened group in the change in BMI
for age z-score between 0 and 18 months. The secondary outcomes are the waist-to-height ratio (%), the sum of the four
skinfolds (mm) and body fat percentage estimated from electrical impedance (kg and %). Waist-to-height ratio is an accurate measure of body fat in children that does not require sex
and age specific adjustments [50]. Skinfold thickness is also
an accurate indicator of body fat in children [51]. We will
estimate fat percentage from the electrical impedance as
follows [52]: Body fat (kg) = body weight fat free mass.
Fat free mass (kg) = 0.622 height 2/R50 (Ohm) + 0.234
weight (kg) + 1.166 R50 = resistance at 50 kHz.

253

2.13. Data analysis and statistics

We consider a per-protocol analysis most suitable because our research question examines the biological effect
of sugar-free beverages on weight instead of the effectiveness
to implement an intervention of sugar-free beverages. We
will also do an intention-to-treat analysis to take into account
selective dropout and to prevent ignoring children who
stopped due to effects of the study beverages on body weight.
Our primary analysis will thus be per-protocol analyses of
those children who consumed the study beverages until the
end of the study. The primary endpoint is BMI for age zscore, and the secondary endpoints are sum of four skinfolds
(mm), waist to height ratio (%) and fat mass determined
from electrical impedance (kg and %). The difference in
change from 0 to 18 months between the sugar-free and
sugar-sweetened group will be analyzed with the independent samples t-test. For each endpoint we will report the
mean difference in change between the sugar-free and
sugar group and its 95% confidence interval and p-value. To
check the robustness of this analysis we will also perform a
linear regression analysis that corrects for the outcome variables at baseline. We do not expect that this will lead to
markedly different results because we randomized the participants into two large groups and stratified for school, gender, age and initial BMI.
Approximately a third of the participants are siblings. We
will perform subgroup-analyses in which we will collapse
siblings from the same household into one fictional subject
who will be assigned the mean value of these siblings. This
will reduce the number of subjects and eliminate effects of
interdependence between siblings.
We will also perform subgroup-analyses for the 80% most
compliant children, where we exclude the 20% of the subjects
with the lowest empty can counts.
Secondary analyses will be intention-to-treat analyses.
These analyses will include children who stopped drinking
the lemonades prematurely but were available for the final
measurements at 18 months. We expect that final measurements will be available for 613 out of 642 children randomized. We will not use the last observation carried forward
method or imputation techniques because there are many
objections against these methods, especially in growing children. We will perform the same analyses on this dataset as
described above for the primary analyses.
The second and third timepoints, t = 6 and t = 12 months,
will be used for graphs but not for analyses. The level of significance will be P b 0.05 (2-tailed). Treatment codes will be broken after blinded data-analysis. For the blinded data-analysis
we will use the numbers 0 and 1 for the treatment groups.
After the analyses we will ask our statistician (LDJK) to deblind the dataset. We will use the Statistical Package for the Social Sciences (SPSS) version 17.0 to perform the analyses.
2.14. Timelines
Fig. 3 presents the planning of the study. After the recruitment and baseline measurements we randomized the children
into the two treatment groups. There were no dropouts
between the baseline measurements and randomization. Children started the study with a run-in period of one week in

254

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

Fig. 3. Study planning of the Double-blind Randomized INtervention study in Kids (DRINK). Body measurements and sensory evaluations are performed at t = 0,
6, 12 and 18 months, dental health questionnaires at t = 12 and 18 months.

which they were all given the sugar-sweetened peach lemonade. This run-in period served to iron out remaining logistic
problems and did not contribute to the length of the intervention period. The first two schools started the intervention on
November 14, 2009, another three schools on November 30,
2009 and the final three schools on December 7, 2009. The
last three schools started without a run-in period because of
time restrictions. We interrupted treatment during the summer holidays of 6 weeks from July 10, 2010August 22, 2010.
Final measurements are done after 18 months intervention
and the study therefore ended in July, 2011.
3. Discussion
Low-caloric drinks seem a healthy alternative to sugarsweetened beverages to prevent obesity in children because
sugar-sweetened drinks may fail to satiate and therefore
add calories on top of the rest of the diet. However, conclusive evidence that sugar-sweetened beverages are more
prone than other foods to cause overweight is lacking. Therefore we designed the DRINK study. Our study examines if
children who are switched to sugar-free beverages increase
their caloric intake from other sources to compensate the
caloric deficit in these beverages, or not.
3.1. Strengths and limitations
Our study has several strengths. First, the study is doubleblind; nobody knows if a child is drinking sugar-sweetened
or sugar-free beverages. Therefore children in the sugar-free
group receive no explicit or subconscious cues that they are
expected to lose weight, and children in the sugary group
are not encouraged to eat less or exercise more so as to
avoid weight gain. Hence this study will be a strict test of

effects of liquid sugars on body weight through unconscious

physiological mechanisms that regulate food intake. Previous
trials in children [2830] and in adults [27] on the relation
between sugar-sweetened beverages and body weight investigated the effects of educational programs or environmental
interventions. These trials were not blinded and could not
separate changes in body weight due to behavioral changes
from changes due to physiological factors. Other trials in
adults [2426] on the relation between sugar-sweetened
beverages and body weight were also not properly blinded.
A second strength is that treatment lasts 18 months and
that we have 642 participants. Other trials [2426] were relatively short, varying from 3 to 10 weeks, and sample sizes
ranged from 15 to 42 subjects.
A third strength is the use of a proper placebo treatment.
A study on the effect of sugar-sweetened beverages on
weight gain requires sugar-sweetened and sugar-free drinks
that taste and look the same. We managed to develop and
produce these beverages. Earlier trials used other treatments.
Tordoff [25] used high-fructose corn syrup versus aspartame
sweetened beverages, Raben [26] used sucrose-sweetened
drinks and foods versus artificially-sweetened drinks and
foods, DiMeglio [24] used jelly beans versus caffeine-free
soda.
The fourth strength is the individual randomization,
which will average out other behaviors that affect weight.
For example, both groups should have similar numbers of
children that watch a lot of television or change their TV
watching habits, that eat a lot of candy, or that become
more or less active. Also, parental education and ethnicity
are similar for both experimental conditions. Furthermore,
we stratified participants by school, gender, age and BMI category to ensure that these potential determinants of weight
gain were evenly distributed between the treatment and

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

control group. Therefore we assume that any difference in

change in endpoints can be attributed to the study beverages
and not to other life style factors.
Our study also has limitations. First, the hypothesis that
we test deals with satiation and food intake, and ideally we
should measure what the children eat and drink in response
to the treatments. However, it is impossible to measure food
intake in free living subjects to the degree of precision required for our study. As body fatness is the outcome of concern, we decided to concentrate on this endpoint. Follow-up
experiments with smaller groups that examine the underlying mechanisms of the association between liquid carbohydrates and weight status might consider measurements of
caloric intake.
Second, the ideal endpoint would have been fat mass estimated from dual-energy X-ray scans. This was not feasible.
However, if the study beverages cause differences in BMI
for age z-score it is plausible that these are due to differences
in fat mass, and we have other body measurements to back
this up.
Third, our primary compliance marker is counts of
returned empty cans that children consumed during school
days. We did not collect empty cans from the weekends
and holidays. In the sugar-free group analysis of sucralose
in urine will provide a measure of compliance that is not
available for the children receiving sugar-sweetened lemonade. However, our objective measurements of compliance
still exceed those in most other trials of diet and body weight.
Fourth, we excluded schools that were difficult to reach or
had few prospective participants. We do not think that this
has an effect on the generalizability of the results. We
recruited schools from an urbanized area near Amsterdam.
We included the children from six towns and excluded children from six towns that partly overlapped. Based on zip
codes of schools included and excluded, socioeconomic status is quite similar and close to the average Dutch socioeconomic status. In addition we do not study behavior but a
physiological mechanism that in principle should not be
much sensitive to socioeconomic status or place of residence.
The study population consists of healthy children and focuses on biological satiation mechanisms that may be widely
shared. We therefore assume that findings of this study can
be extrapolated beyond the healthy Dutch children of our
study. On the other hand the effect of sugar-sweetened
drinks on body weight might depend on body size, genetics,
ethnic factors or health, and thus our findings should be extrapolated with caution and confirmed in other populations.

255

would help to prevent weight gain because they are a major

source of calories.
If both treatments have the same effect on weight it could
also be argued that the artificial sweeteners in the sugar-free
beverages stimulated caloric intake [53]. Some animal studies
supported this hypothesis [5456] but experiments in humans
[5759] did not. The hypothesis that artificial sweeteners increase energy intake would therefore not be the most likely explanation if we find no difference between the groups.
If we find that weight gain slows after removal of liquid
sugar then children in the sugar-free group evidently did
not increase their intake of other foods and drinks sufficiently
to compensate their reduced intake of liquid calories. It
would also suggest that the converse holds true, i.e. that adding liquid sugar to the diet causes weight gain because it is
not spontaneously compensated by reduced intake of other
foods. That would corroborate the liquid sugar theory and
suggest a unique role for liquid sugars in causing weight
gain in children. This outcome would further support current
recommendation to discourage sugar-sweetened drinks and
promote non-caloric drinks such as water to prevent weight
gain in children. However, one piece of evidence would still
be missing: does covert removal of calories from solid foods
also lead to weight reduction? Current theory suggests that
it would not, and that such calories would be compensated
for from other foods. However, only a double blind trial similar to ours can provide decisive evidence on this reduction.
Conict of interests
The authors declare that they have no conflict of interest.
Author contributions
J.C.R. coordinated the pilot study, sensory study and the
main trial. She also supervised the data collection, coobtained medical ethical approval, supervised data management, was responsible for the recruitment of subjects, managed logistics of the trial and drafted the manuscript. M.R.O.
co-obtained funding, obtained medical ethical approval, contributed to the design, co-supervised the study and cosupervised the development and manufacture of the beverages. L.D.J.K. randomized the participants and performed
the coding of the beverages. M.B.K supervised the study, conceived and designed the study, obtained funding and supervised the development and manufacture of the beverages.
M.R.O, L.D.J.K and M.B.K. revised the manuscript critically.
Acknowledgments

3.2. Interpretation of outcomes

There are two possible outcomes of our study. If we find no
difference in weight gain between the sugar-free and sugarsweetened group, then children in the sugar-free group must
have increased their caloric intake from other sources to
compensate for the calories lacking from the study beverage.
An exclusive focus on sugar-sweetened beverages might then
be less effective to decrease overweight in children. Other interventions than reducing intake of sugar-sweetened beverages to decrease overweight in children should receive equal
emphasis. Still, a lower intake of sugar-sweetened beverages

We thank the staff, teachers, parents and children of the

schools for their willingness to participate and the pleasant
cooperation during the study; Emilie de Zoete and Hetty
Geerars for excellent assistance in the execution of this
study; Mr. Joop Bremer, retired head of research of a Dutch
soft drink manufacturer, for his help and advice during development and manufacture of the beverages; Refresco Benelux
for meticulous care in producing the beverages; Prof. dr.
Jacob Seidell for valuable advice during the study; the members of our advisory board: Sera de Vries, teacher, Edgar van
Mil, MD, Phd, pediatricendocrinologist and Goof Buijs, M.Sc.,

256

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

senior advisor at Netherlands Institute for Health Promotion

and Disease Prevention for creative suggestions; Herbert van
den Heuvel and colleagues, for their dedication in the beverage
distribution; Ellinore Tellegen, for acting as independent
physician.
References
[1] Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 19992004. JAMA
2006;295:154955.
[2] Ogden CL, Carroll MD, Curtin LR, Lamb MM, Flegal KM. Prevalence of
high body mass index in US children and adolescents, 20072008.
JAMA 2010;303:2429.
[3] Hayman LL, Meininger JC, Daniels SR, McCrindle BW, Helden L, Ross J,
et al. Primary prevention of cardiovascular disease in nursing practice:
focus on children and youth: a scientific statement from the American
Heart Association Committee on Atherosclerosis, Hypertension, and
Obesity in Youth of the Council on Cardiovascular Disease in the
Young, Council on Cardiovascular Nursing, Council on Epidemiology
and Prevention, and Council on Nutrition, Physical Activity, and Metabolism. Circulation 2007;116:34457.
[4] Hannon TS, Rao G, Arslanian SA. Childhood obesity and type 2 diabetes
mellitus. Pediatrics 2005;116:47380.
[5] Steinberger J, Daniels SR. Obesity, insulin resistance, diabetes, and cardiovascular risk in children: an American Heart Association scientific
statement from the Atherosclerosis, Hypertension, and Obesity in the
Young Committee (Council on Cardiovascular Disease in the Young)
and the Diabetes Committee (Council on Nutrition, Physical Activity,
and Metabolism). Circulation 2003;107:144853.
[6] Magarey AM, Daniels LA, Boulton TJ, Cockington RA. Predicting obesity
in early adulthood from childhood and parental obesity. Int J Obes Relat
Metab Disord 2003;27:50513.
[7] Lobstein T, Baur L, Uauy R. Obesity in children and young people: a crisis in public health. Obes Rev 2004;5(Suppl. 1):482.
[8] Franks PW, Hanson RL, Knowler WC, Sievers ML, Bennett PH, Looker
HC. Childhood obesity, other cardiovascular risk factors, and premature
death. N Engl J Med 2010;362:48593.
[9] Weiss R, Dziura J, Burgert TS, Tamborlane WV, Taksali SE, Yeckel CW,
et al. Obesity and the metabolic syndrome in children and adolescents.
N Engl J Med 2004;350:236274.
[10] Duffey KJ, Popkin BM. Shifts in patterns and consumption of beverages
between 1965 and 2002. Obesity 2007;15:273947 (Silver Spring).
[11] Nielsen SJ, Popkin BM. Changes in beverage intake between 1977 and
2001. Am J Prev Med 2004;27:20510.
[12] Mattes R. Fluid calories and energy balance: the good, the bad, and the
uncertain. Physiol Behav 2006;89:6670.
[13] Schulze MB, Manson JE, Ludwig DS, Colditz GA, Stampfer MJ, Willett WC,
et al. Sugar-sweetened beverages, weight gain, and incidence of type 2
diabetes in young and middle-aged women. JAMA 2004;292:92734.
[14] Palmer JR, Boggs DA, Krishnan S, Hu FB, Singer M, Rosenberg L. Sugarsweetened beverages and incidence of type 2 diabetes mellitus in
African American women. Arch Intern Med 2008;168:148792.
[15] Bes-Rastrollo M, Sanchez-Villegas A, Gomez-Gracia E, Martinez JA,
Pajares RM, Martinez-Gonzalez MA. Predictors of weight gain in a
Mediterranean cohort: the Seguimiento Universidad de Navarra Study
1. Am J Clin Nutr 2006;83:36270 quiz 945.
[16] Nicklas TA, Yang SJ, Baranowski T, Zakeri I, Berenson G. Eating patterns
and obesity in children. The Bogalusa Heart Study. Am J Prev Med
2003;25:916.
[17] Ludwig DS, Peterson KE, Gortmaker SL. Relation between consumption
of sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. Lancet 2001;357:5058.
[18] Dubois L, Farmer A, Girard M, Peterson K. Regular sugar-sweetened
beverage consumption between meals increases risk of overweight
among preschool-aged children. J Am Diet Assoc 2007;107:92434.
[19] Johnson L, Mander AP, Jones LR, Emmett PM, Jebb SA. Is sugarsweetened beverage consumption associated with increased fatness
in children? Nutrition 2007;23:55763.
[20] Newby PK, Peterson KE, Berkey CS, Leppert J, Willett WC, Colditz GA.
Beverage consumption is not associated with changes in weight and
body mass index among low-income preschool children in North Dakota.
J Am Diet Assoc 2004;104:108694.
[21] O'Connor TM, Yang SJ, Nicklas TA. Beverage intake among preschool
children and its effect on weight status. Pediatrics 2006;118:e10108.
[22] Vandenbroucke JP. Observational research, randomised trials, and two
views of medical science. PLoS Med 2008;5:33943.

[23] Fewell Z, Davey Smith G, Sterne JAC. The impact of residual and unmeasured confounding in epidemiologic studies: a simulation study. Am J
Epidemiol 2007;166:64655.
[24] DiMeglio DP, Mattes RD. Liquid versus solid carbohydrate: effects on
food intake and body weight. Int J Obes Relat Metab Disord 2000;24:
794800.
[25] Tordoff MG, Alleva AM. Effect of drinking soda sweetened with aspartame or high-fructose corn syrup on food intake and body weight.
Am J Clin Nutr 1990;51:9639.
[26] Raben A, Vasilaras TH, Moller AC, Astrup A. Sucrose compared with artificial sweeteners: different effects on ad libitum food intake and body
weight after 10 wk of supplementation in overweight subjects. Am J
Clin Nutr 2002;76:7219.
[27] Chen L, Appel LJ, Loria C, Lin PH, Champagne CM, Elmer PJ, et al. Reduction in consumption of sugar-sweetened beverages is associated with
weight loss: the PREMIER trial. Am J Clin Nutr 2009;89:1299306.
[28] James J, Thomas P, Cavan D, Kerr D. Preventing childhood obesity by
reducing consumption of carbonated drinks: cluster randomised controlled trial. BMJ 2004;328:1237.
[29] Ebbeling CB, Feldman HA, Osganian SK, Chomitz VR, Ellenbogen SJ, Ludwig
DS. Effects of decreasing sugar-sweetened beverage consumption on body
weight in adolescents: a randomized, controlled pilot study. Pediatrics
2006;117:67380.
[30] Sichieri R, Paula Trotte A, de Souza RA, Veiga GV. School randomised
trial on prevention of excessive weight gain by discouraging students
from drinking sodas. Public Health Nutr 2009;12:197202.
[31] Katan MB. Weight-loss diets for the prevention and treatment of obesity.
N Engl J Med 2009;360:9235.
[32] Crawford PB, Woodward-Lopez G, Ritchie L, Webb K. How discretionary can we be with sweetened beverages for children? J Am Diet
Assoc 2008;108:14404.
[33] Allison DB, Mattes RD. Nutritively sweetened beverage consumption
and obesity: the need for solid evidence on a fluid issue. JAMA
2009;301:31820.
[34] SCF. Sucralose (opinion adapted by the SCF on 7 September 2000).
http://ec.europa.eu/food/fs/sc/scf/out68_en.pdfFebruary 19, 2011.
[35] SCF. Re-evaluation of acesulfame K with reference to the previous SCF
opinion of 1991 (opinion expressed on 9th March 2000). http://ec.
europa.eu/food/fs/sc/scf/out52_en.pdfFebruary 19, 2011.
[36] Snedecor GW, Cochran WG. Statistical methods. 8th ed. Ames, USA:
Iowa State University Press; 1989.
[37] Federal Register. National Bureau of Standards. Customary system of
weight and measures. Fed Register 1968;33:345.
[38] Katan MB, Ludwig DS. Extra calories cause weight gainbut how much?
Jama 2010 Jan 6;303:656.
[39] Wang YC, Gortmaker SL, Sobol AM, Kuntz KM. Estimating the energy
gap among US children: a counterfactual approach. Pediatrics
2006;118:e172133.
[40] Fredriks AM, van Buuren S, Burgmeijer RJ, Meulmeester JF, Beuker RJ,
Brugman E, et al. Continuing positive secular growth change in The
Netherlands 19551997. Pediatr Res 2000;47:31623.
[41] de Jong E, Schokker DF, Visscher TL, Seidell JC, Renders CM. Behavioural
and socio-demographic characteristics of Dutch neighbourhoods with high
prevalence of childhood obesity. Int J Pediatr Obes 2011 Aug 6;6:298305.
[42] Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey.
BMJ 2000;320:12403.
[43] Cole TJ, Flegal KM, Nicholls D, Jackson AA. Body mass index cut offs to
define thinness in children and adolescents: international survey. BMJ
2007;28:1667.
[44] CBS. Number of non-western people in the Netherlands. Netherlands.
http://statline.cbs.nl/StatWeb/publication/?DM=SLNL&PA=70787NED&D
1=a&D2=0&D3=0&D4=a&D5=l&HDR=T,G3&STB=G1,G2,G4&VW=T
February 19, 2011.
[45] CBS. Level of education of the Netherlands population (1565 year olds).
Netherlands. http://statline.cbs.nl/StatWeb/publication/?DM=SLNL&PA=
71822NED&D1=0&D2=0&D3=0&D4=0&D5=0-2,4-10&D6=0&D7=l&
HDR=T,G1&STB=G2,G4,G5,G3,G6&VW=T February 19, 2011.
[46] Schulz KF, Grimes DA. Generation of allocation sequences in randomised trials: chance, not choice. Lancet 2002;359:5159.
[47] Gibson RS. Principles of nutritional assessment. 2nd ed. New York, USA:
Oxford University Press; 2005.
[48] Heymsfield SB, Lohman TG, Wang Z, Going SB. Human body composition 2nd ed. Champaign, USA: Human Kinetics; 2005.
[49] Roberts A, Renwick AG, Sims J, Snodin DJ. Sucralose metabolism and
pharmacokinetics in man. Food Chem Toxicol 2000;38(Suppl. 2):
S3141.
[50] Garnett SP, Baur LA, Cowell CT. Waist-to-height ratio: a simple option
for determining excess central adiposity in young people. Int J Obes
(Lond) 2008;32:102830.

J.C. de Ruyter et al. / Contemporary Clinical Trials 33 (2012) 247257

[51] Deurenberg P, Pieters JJ, Hautvast JG. The assessment of the body fat
percentage by skinfold thickness measurements in childhood and
young adolescence. Br J Nutr 1990;63:293303.
[52] Rush EC, Puniani K, Valencia ME, Davies PS, Plank LD. Estimation of
body fatness from body mass index and bioelectrical impedance: comparison of New Zealand European, Maori and Pacific Island children.
Eur J Clin Nutr 2003;57:1394401.
[53] Rosenman K. Benefits of saccharin: a review. Environ Res 1978;15:
7081.
[54] Davidson TL, Swithers SE. A Pavlovian approach to the problem of obesity. Int J Obes Relat Metab Disord 2004;28:9335.
[55] Swithers SE, Baker CR, Davidson TL. General and persistent effects of
high-intensity sweeteners on body weight gain and caloric compensation in rats. Behav Neurosci 2009;123:77280.

257

[56] Swithers SE, Davidson TL. A role for sweet taste: calorie predictive relations in energy regulation by rats. Behav Neurosci 2008;122:16173.
[57] Rodin J. Comparative effects of fructose, aspartame, glucose, and water
preloads on calorie and macronutrient intake. Am J Clin Nutr 1990;51:
42835.
[58] Black RM, Tanaka P, Leiter LA, Anderson GH. Soft drinks with aspartame: effect on subjective hunger, food selection, and food intake of
young adult males. Physiol Behav 1991;49:80310.
[59] Ford HE, Peters V, Martin NM, Sleeth ML, Ghatei MA, Frost GS, et al.
Effects of oral ingestion of sucralose on gut hormone response and
appetite in healthy normal-weight subjects. Eur J Clin Nutr 2011;65:
50813.