European Journal of Clinical Nutrition

https://doi.org/10.1038/s41430-017-0006-9

ARTICLE

Effect of high milk and sugar-sweetened and non-caloric soft

drink intake on insulin sensitivity after 6 months in overweight
and obese adults: a randomized controlled trial
Sara Engel1 Tine Tholstrup1 Jens M Bruun1,2,3 Arne Astrup1 Bjørn Richelsen2 Anne Raben1
● ● ● ● ●

Received: 7 June 2017 / Revised: 4 August 2017 / Accepted: 24 August 2017

© Macmillan Publishers Limited, part of Springer Nature 2017

Abstract
Background/objectives Milk contributes with saturated fat, but randomized controlled trials (RCT) on the effects of dairy
on the risk of type 2 diabetes (T2D) where dairy is given as whole foods are scarce. The objective of our study was to
investigate the long-term effects of semi-skimmed milk on insulin sensitivity and further to compare milk with sugar-
sweetened soft drinks (SSSD).
Subjects/methods A secondary analysis of a 6-month RCT with 60 overweight and obese subjects randomly assigned to 1
L/d of either milk (1.5 g fat/100 mL), SSSD, non-calorie soft drink (NCSD), or water was conducted. Insulin sensitivity was
evaluated by oral glucose tolerance test (OGTT) and plasma free fatty acids. Second, fasting blood lipids, blood pressure,
and concentration of plasminogen activator inhibitor-1 were assessed.
Results There were no differences between milk, SSSD, NCSD, and water on insulin sensitivity assessed by OGTT
(Matsuda Index, fasting, and area under the curve glucose, insulin and homeostasis model assessment values). SSSD
increased total cholesterol compared to NCSD (P = 0.007), and triacylglycerol compared to NCSD and water (P = 0.045
and 0.045, respectively). None of the other parameters differed significantly between the groups.
Conclusions In conclusion, there were no differences in effect between intake of milk, SSSD, NCSD, and water (1 L/d) for
6-month on risk markers of T2D in overweight and obese adults. As a secondary analysis, these results need confirmation in
future studies.

Introduction low-density lipoprotein (LDL) cholesterol. However, meta-

Type 2 diabetes (T2D) is on the rise and currently accounts
for the majority of cases of diabetes worldwide [1]. Keeping
a healthy diet can prevent or delay the onset of T2D [1].
Today, dietary guidelines recommend intake of low-fat dairy
products (milk<0.5% fat) because of the high content of
saturated fatty acids (SFAs), which is known to increase
* Sara Engel
saraengel@nexs.ku.dk
1 ficial increase in fasting insulin with milk compared to a
Department of Nutrition, Exercise and Sports, Faculty of Science,
University of Copenhagen, Rolighedsvej 30, DK-1958
Frederiksberg, Denmark
2
Department of Endocrinology and Internal Medicine (MEA),
Aarhus University Hospital, Aarhus, Denmark
3
Department of Internal Medicine, Randers Regional Hospital,
Randers, Denmark
analyses of observational studies indicate that intake of dairy
is not associated with increased risk of cardiovascular dis-
ease (CVD) [2, 3] and point to an inverse association
between overall dairy intake and T2D [4–6]. Milk con-
stitutes a considerable part of dairy consumption, but only
few studies have exclusively examined the effect of milk
intake on insulin and glucose, and results have been
inconclusive. Two studies of 1 year and 6 weeks, respec-
tively, found no effect on fasting insulin and glucose com-
paring semi-skimmed milk to skimmed milk or to a non-
dairy diet [7, 8]; one study reported a significant, unbene-
non-dairy diet after 2 weeks [9], and another found a sig-
nificant, unbeneficial increase in glucose concentration in
healthy subjects with milk compared to a low-dairy diet after
12 weeks [10]. Sugar-sweetened soft drinks (SSSD) are a
popular but questionable alternative to milk. Thus, research
show that SSSD can be associated with overweight, T2D,

Fig. 1 Participant flowchart. NCSD non-caloric soft drink, SSSD sugar-sweetened soft drink
and the metabolic syndrome [11–13]. Additionally, we
reported in the primary study that SSSD increased ectopic
fat accumulation and elevated concentration of uric acid in
overweight and obese subjects [14, 15]. A comparison
between milk and the somewhat controversial beverages
SSSD and artificially sweetened non-caloric soft drinks
(NCSD) are highly applicable to everyday choices for the
normal population. The aim of this randomized controlled
trial (RCT) was therefore to compare the effect of 1 L/d of
semi-skimmed milk (MILK) (1.5 g fat/100 mL) with SSSD
and NCSD on insulin sensitivity after 6 months of intake,
measured by a 2-h oral glucose tolerance test (OGTT), and
concentration of free fatty acids (FFA). The 2-h OGTT is a
valid and recognized method to evaluate whole-body glu-
cose tolerance and insulin sensitivity [16, 17]. Water
(WATER) was included as a non-caloric control. Second,
risk markers of CVD; fasting lipids, plasminogen activator
inhibitor-1 (PAI-1), and blood pressure were reported. We
hypothesized that consumption of MILK would favorably
affect insulin sensitivity and risk markers of CVD in over-
weight and obese subjects compared with isocaloric amounts
of SSSD, NCSD, and WATER.
Materials and methods
Subjects
Sixty healthy, non-diabetic subjects were recruited from two
study sites at Aarhus University Hospital and University of
Copenhagen. A total of 90 individuals were assessed for
S. Engel et al.
eligibility, 73 were recruited, and 13 (all women) dropped
out after randomization (Fig. 1). The primary study was
described in detail elsewhere [14, 15]; however, with fewer
participants included (47). Due to logistical problems in one
of the centers (pregnancy and illness) the remaining 13
participants (9 women and 4 men) were included later. The
following end points included in the present study have
therefore previously been published with fewer participants;
fasting glucose, insulin, blood lipids, and blood pressure
[14]. The most important advantage of this study was the
inclusion of OGTT and plasma concentration of FFA and
PAI-1 enabling us to study the effect of MILK compared to
SSSD on insulin sensitivity, not reported in the primary
study. Thus, the outcome was secondary and considered
explorative, as no pre-study power calculations were per-
formed. Inclusion criteria were body mass index (BMI)
(26–40 in kg/m2) and age (20–50 years). Exclusion criteria
were as follows: diabetes, hypertension (>160/100 mmHg),
medication affecting either blood lipids, blood glucose or
body weight, smoking, pregnancy or breastfeeding, aller-
gies to milk or suffering from phenylketonuria, and exces-
sive physical activity (>10 h/week). All subjects gave their
informed consent in writing. The study was done in
agreement with the Helsinki Declaration and approved by
the Ethics Committee of Middle Jutland, Denmark.
Intervention
The composition of the test beverages are shown in Table 1.
Using block randomization, subjects were allocated to
consume either 1 of 4 test beverages for 6 months: semi-
Milk and insulin sensitivity

Table 1 Composition of the test beveragesa Hospital) carried out all analyses. Blood samples for the
Milk SSSD NCSD Water remaining 13 study subjects not included in the primary
study [14] were analyzed in 2015 using the exact same
Carbohydrate (g/100 mL) 4.7 10.6 0 0
routines and methods. FFA and PAI-1 were not analyzed
Protein (g/100 mL) 3.4 0 <0.1 0 for the primary study, thus analyses for all 60 subjects were
Fat (g/100 mL) 1.5 0 0 0 done in 2015. Blood pressure was recorded with a digital
SFA 1.0 0 0 0 blood pressure apparatus (Colin Press-Mate) after a 10 min
MUFA 0.4 0 0 0 rest. To evaluate insulin sensitivity, the Matsuda Index was
PUFA 0.1 0 0 0 calculated including values for glucose and insulin at all the
Energy (kJ/d) 1900 1800 15 0 time points during the OGTT and integrated in the fol-
Volume (mL) 1000 1000 1000 1000 lowing formula [16]: 10,000/√((fasting plasma glucose ×
MUFA monounsaturated fatty acid, NCSD non-calorie soft drink, fasting plasma insulin) × (mean OGTT glucose × mean
PUFA polyunsaturated fatty acid, SSSD sugar-sweetened soft drink, OGTT insulin)). In addition, insulin resistance was eval-
SFA saturated fatty acid uated by using HOMA with the following formula [18]:
a
Subjects were randomized to consume 1 L of milk, SSSD, NCSD, or fasting serum insulin (µU/mL) × fasting plasma glucose
WATER daily for 6 months (mmol/L)/22.5. Body composition was determined by DXA
(Hologic 2000/W QDR scanner; Hologic Inc) before and
skimmed milk (1.5%), MILK (Arla Foods, Denmark), after intervention with determination of fat mass (FM) and
sucrose-sweetened regular cola (50% glucose and 50% fat-free mass. The DXA output was generated with default
fructose), SSSD and aspartame-sweetened diet cola, NCSD settings of the Hologic software version 12.6.1.
(Coca Cola, Denmark) or still mineral water, WATER
(Aqua d’or, Brande, Denmark). The study was not blinded Statistical analysis
due to the nature of the intervention. All test beverages were
free of charge and subjects were instructed to consume 1 L/ In this secondary analysis, the effects of four beverages on
day and allowed to drink water, coffee, tea, and their regular insulin and glucose evaluated by OGTT at baseline and
amount of alcohol. Test beverages were handed out 2–3 after 6 months were assessed by using a linear mixed ana-
times/month from the research center and all subjects were lysis of covariance (ANCOVA) model including interaction
instructed to bring back empty bottles/cartons for monitor- between time and treatment to assess if time modified the
ing and compliance. The lifestyle changes concerning diet effect of the beverage groups during the time of the 2-h
and physical activity were monitored at baseline and end of OGTT. An approximate F-test was used to evaluate the
the intervention using a 7-day weighed dietary record interaction between time and treatment and if non-
(Dankost Pro dietary assessment software (Dankost)) and a significant another approximate F-test was used to evalu-
validated questionnaire about physical activity [14]. ate if there was a time-independent treatment effect. Ran-
dom effects were included to account for between-subject
Clinical investigations variation and to account for year of analysis of blood
samples (2011/2015). The time course of the estimated
Subjects consumed a standardized evening meal and fasted treatment effect for all four beverages was shown graphi-
overnight. The following morning blood samples were cally with the corresponding standard errors. Other outcome
collected (at 0800 hours) and a 2-h OGTT was performed. variables were analyzed using an ANCOVA model
Blood was collected at 0, 30, 60, and 120 min but for insulin including the treatment. The treatment effect was evaluated
the concentration at 30 min was not measured. The subjects by means of an F-test. To adjust for differences between
were instructed to refrain from medicine, alcohol, and vig- beverages and to take biological variation into account,
orous exercise for 24 h before blood sampling. Total and baseline values were included as well as covariates; age,
high-density lipoprotein (HDL) cholesterol, triacylglycerol, gender, baseline BMI, and change in FM (kg). All models
and glucose were analyzed using routine laboratory meth- were also adjusted for year of analysis of blood samples
ods at the hospital (Department of Clinical Biochemistry, (2011/2015) besides models for blood pressure, PAI-1, and
Aarhus University Hospital). LDL cholesterol was calcu- FFA (all 60 analyzed at the same time). For dietary records
lated by the Friedewald equation. Commercially available and anthropometric outcomes, statistical differences were
kits were used to assess insulin (Human Insulin ELISA; also analyzed using ANCOVA model with baseline and
DAKO, Glostrup, Denmark). FFAs were assessed by using gender (only anthropometric) as covariates. If a model
an enzymatic colorimetric method assay (Wako showed treatment effect, pairwise comparisons between the
Diagnostics-NEFA (Free Fatty Acid). The same laboratory four beverages were subsequently done with Tukey method
(Department of Clinical Biochemistry, Aarhus University for multiple comparisons. Treatment differences were
 S. Engel et al.

Table 2 Sex, age,

Milk SSSD NCSD Water P-value
anthropometric measurements,
and body composition at Sex (n (%)) 15 (25) 14 (23.3) 15 (25) 16 (26.7)
baseline, after 6 months
intervention, and differences Women 11 (18.3) 6 (10) 12 (20) 11 (18.3)
between the beverage groupsa Men 4 (6.7) 8 (13.3) 3 (5) 5 (8.3)
Age, yb 37.7 ± 9.1 37.8 ± 8.0 39 ± 7.6 39 ± 7.3 0.94
Body weight, kg
Pre-intervention 94.0 ± 17.1 94.9 ± 11.6 94.5 ± 12.6 98.4 ± 21.8 0.87
Post-intervention 95.5 ± 4.3 96.4 ± 3.0 95.0 ± 3.3 99.1 ± 5.8 0.43
BMI, kg/m2
Pre-intervention 31.4 ± 3.0 30.8 ± 2.8 33.4 ± 4.1 31.5 ± 4.5 0.25
Post-intervention 32.0 ± 0.8 31.3 ± 0.7 33.6 ± 1.1 31.7 ± 1.1 0.30
Fat mass, kg
Pre-intervention 34.6 ± 6.9 33.5 ± 7.5 38.3 ± 11.3 35.7 ± 10.9 0.56
Post-intervention 35.9 ± 2.4 35.6 ± 2.3 38.5 ± 2.8 35.9 ± 3.3 0.27
Fat-free mass, kg
Pre-intervention 56.7 ± 14.1 59.7 ± 11.3 53.6 ± 10.6 56.3 ± 12.3 0.64
Post-intervention 57.3 ± 3.8 58.4 ± 3.1 54.1 ± 2.8 56.4 ± 3.3 0.34
The subjects were randomly assigned to the four groups of 1 L daily of test beverage
NCSD non-caloric soft drink, SSSD sugar-sweetened soft drink
a
All values are means ± SEs. Statistical differences were analyzed in an ANCOVA model with gender and
values from baseline included as covariates. n = 60 (n = 58 for fat mass and fat-free mass due to two subjects
had missing values)
b
For age, values are means ± SDs and the difference was analyzed by ANOVA

reported in terms of unadjusted mean levels with corre-
sponding standard errors. All models were validated by
graphical assessment of normal quantile plots and residual
vs. fitted plots. When departures were detected, the vari-
ables were transformed using the logarithm transformation.
We tested for variance homogeneity and the test showed
similar variance. A two-tailed P-value < 0.05 was con-
sidered significant. The statistical software R version 3.1.3
2015 was used for all statistical evaluations.
Results
Insulin and glucose changes
Analysis of the OGTT measured before and after the
intervention showed no significant time× treatment effect
or treatment effect during the 2 h test for glucose (P = 0.601
and 0.835, respectively) or insulin (P = 0.349 and 0.552,
respectively) (Fig. 2a, b). Calculated Matsuda Index and
AUC for glucose and insulin based on OGTT data, fasting
glucose and insulin, as well as calculated HOMA values
based on fasting and AUC for glucose and insulin are all
listed in Table 3. In line with the OGTT test results, there
were no differences between the groups.

Subjects Blood lipids

Of the 73 randomly assigned subjects, 60 completed the
6 months intervention. Five subjects in the SSSD group,
four subjects in the NCSD group, four subjects in the
WATER group, and none in the MILK group dropped out
(all women). Baseline characteristics of the 60 subjects who
completed the study are listed in Tables 2 and 3. Results
showed no difference between the four intervention groups;
however the distribution of gender was unequal and the
models were adjusted accordingly. Anthropometric mea-
sures and body composition pre-intervention and post-
intervention are listed in Table 2. No significant differences
were observed after 6 months.
Results from fasting blood lipid concentrations measured
after 6 months intervention are listed in Table 3. The ana-
lysis showed significant differences between the groups for
total cholesterol and triacylglycerol (P = 0.01 and 0.02,
respectively). For total cholesterol, the pairwise compar-
isons showed that SSSD was statistically higher compared
to NCSD (P = 0.007), whereas there was a tendency for
total cholesterol to be lower after NCSD compared to MILK
and WATER (P = 0.089 and 0.074, respectively). For
triacylglycerol, the pairwise comparisons showed a sig-
nificantly higher concentration with SSSD compared to
NCSD and WATER (P = 0.045 and 0.045, respectively).
Milk and insulin sensitivity

Table 3 OGTT values, fasting values, and blood pressure at baseline, after 6 months intervention, and differences between the beverage groupsa
Milk SSSD NCSD Water P-value

AUC OGTT glucose, mmol/L

Pre-intervention 894 ± 289 889 ± 228 864 ± 214 829 ± 204 0.86
Post-intervention 911 ± 82 883 ± 60 876 ± 61 812 ± 62 0.86
AUC OGTT insulin, pmol/L
Pre-intervention 36,745 ± 33,909 24,834 ± 8808 32,396 ± 12,305 25,784 ± 10,735 0.30
Post-intervention 27,945 ± 6662 24,364 ± 4889 22,766 ± 3090 17,786 ± 2726 0.49
Fasting glucose, mmol/L
Pre-intervention 5.42 ± 0.71 5.48 ± 0.52 5.52 ± 0.47 5.26 ± 0.54 0.59
Post-intervention 5.69 ± 0.27 5.62 ± 0.18 5.49 ± 0.15 5.35 ± 0.15 0.47
Fasting insulin, pmol/L
Pre-intervention 81.68 ± 71.16 55.32 ± 22.88 75.35 ± 30.64 68.09 ± 54.67 0.52
Post-intervention 91.84 ± 19.40 61.10 ± 6.77 74.73 ± 10.13 93.19 ± 24.94 0.73
Matsuda Index
Pre-intervention 8.85 ± 6.48 8.08 ± 4.56 6.01 ± 3.55 7.98 ± 4.07 0.42
Post-intervention 7.88 ± 1.59 6.58 ± 0.74 6.14 ± 0.84 8.38 ± 1.27 0.79
HOMA-IRb
Pre-intervention 1.55 ± 1.36 1.06 ± 0.44 1.44 ± 0.58 1.29 ± 1.04 0.53
Post-intervention 1.76 ± 0.38 1.17 ± 0.13 1.42 ± 0.19 1.71 ± 0.44 0.56
HOMA-IR AUC
Pre-intervention 2.95 ± 2.94 2.12 ± 0.89 2.58 ± 1.50 3.08 ± 3.26 0.53
Post-intervention 3.60 ± 0.98 2.12 ± 0.25 2.58 ± 0.39 3.08 ± 0.82 0.65
PAI-1, ng/mL
Pre-intervention 55.91 ± 37.21 50.48 ± 34.82 75.90 ± 64.34 44.62 ± 43.46 0.29
Post-intervention 46.1 ± 7.8 52.9 ± 13.3 49.4 ± 9.0 41.2 ± 6.6 0.60
FFA, mmol/L
Pre-intervention 0.59 ± 0.19 0.53 ± 0.21 0.63 ± 0.17 0.60 ± 0.22 0.58
Post-intervention 0.50 ± 0.05 0.55 ± 0.05 0.62 ± 0.07 0.54 ± 0.06 0.33
Cholesterol, mmol/L
Total
Pre-intervention 5.04 ± 1.10 4.84 ± 0.84 5.45 ± 0.84 5.28 ± 0.75 0.28
Post-intervention 5.17 ± 0.23c,d 5.25 ± 0.28d 4.95 ± 0.18c 5.30 ± 0.19c,d 0.01
LDL
Pre-intervention 3.17 ± 0.95 3.07 ± 0.82 3.50 ± 0.98 3.38 ± 0.67 0.53
Post-intervention 3.21 ± 0.20 3.32 ± 0.23 3.17 ± 0.18 3.48 ± 0.17 0.10
HDL
Pre-intervention 1.16 ± 0.26 1.16 ± 0.23 1.18 ± 0.29 1.15 ± 0.29 0.99
Post-intervention 1.23 ± 0.08 1.21 ± 0.09 1.17 ± 0.07 1.21 ± 0.07 0.10
Total:HDL
Pre-intervention 4.54 ± 1.35 4.26 ± 0.80 4.83 ± 1.15 5.09 ± 2.57 0.55
Post-intervention 4.41 ± 0.30 4.49 ± 0.24 4.41 ± 0.28 4.64 ± 0.40 0.49
Triacylglycerol, mmol/L
Pre-intervention 1.58 ± 0.93 1.35 ± 0.73 1.73 ± 0.63 1.67 ± 0.76 0.57
Post-intervention 1.61 ± 0.21c,d 1.60 ± 0.23c 1.37 ± 0.11d 1.36 ± 0.13d 0.02
Blood pressure, mmHg
Systolic
Pre-intervention 124.6 ± 14.9 123.4 ± 8.9 131.5 ± 14.1 124.2 ± 10.9 0.29
Post-intervention 121.2 ± 3.1 125.9 ± 2.9 126.8 ± 2.4 124.1 ± 2.4 0.18
 S. Engel et al.

Table 3 (continued)
Milk SSSD NCSD Water P-value

Diastolic
Pre-intervention 76.0 ± 8.5 74.4 ± 9.7 81.2 ± 7.9 74.9 ± 8.9 0.15
Post-intervention 72.8 ± 2.4 77.5 ± 2.0 77.4 ± 2.3 75.5 ± 2.2 0.22
NCSD non-caloric soft drink, SSSD sugar-sweetened soft drink
a
All values are unadjusted means ± SEs, n = 58, as two subjects had missing values for change in FM
b
n = 57, as one observation was considered an outlier. Statistical differences were analyzed in an ANCOVA model with Tukey pairwise
comparisons adjusted for pre-intervention and the covariates; age, gender, baseline BMI, and change in FM (kg). Values that have no superscript in
common are significantly different from each other (Tukey’s HSD, P < 0.05)

listed in Table 3. There were no differences between the

four beverage groups.

Dietary records

Results from the 7-day weighed dietary records before and

during the intervention are listed in Table 4. No differences
were observed in energy intake between the beverage
groups. Overall, the intake of macronutrients followed the
pattern expected with the different beverages. However, the
intake of fat (percentage of fat) was significantly higher
with WATER compared to MILK and SSSD (P < 0.05 and
P < 0.01, respectively) but not compared to NCSD. When
fat intake was calculated in grams, there were no differences
between the beverage groups. Calcium intake was sig-
nificantly higher with MILK compared to all other groups
before intervention and during intervention (P = 0.01 and P
< 0.001, respectively). A student's t test within the MILK
beverage group showed a significantly higher calcium
intake after 6 months compared to before (t-test: P < 0.001).

Fig. 2 a, b Oral glucose tolerance test (OGTT) measurements per-
formed after 6 months of intervention. All values are means ± SE. n =
58 (a) and 57 (b). Subjects were randomly assigned to the four groups
of 1 L daily of test beverage. Glucose tolerance was measured using a
standard 75 g oral glucose tolerance test. There was no significant
difference in the time course of glucose (a) and insulin (b) con-
centration after the four different beverage groups according to the
repeated measures analysis of variance. NCSD non-caloric soft drink,
SSSD sugar-sweetened soft drink
For LDL, HDL, and total:HDL cholesterol, there were no
differences between the four beverage groups after
6 months.
Blood pressure, PAI-1, and FFA
Results from the blood pressure measurements and con-
centration of PAI-1 and FFA after 6 months intervention are
Discussion
In the present study, we observed that 1 L/d of semi-
skimmed milk (MILK) did not differ from beverages such
as SSSD, NCSD, or WATER in effects on risk markers of
T2D evaluated by OGTT in overweight and obese men and
women after 6 months’ intake. Moreover, the beverages did
not differ in effects on plasma concentrations of PAI-1,
FFA, or blood lipids. Unexpectedly, SSSD showed no
adverse effect on insulin sensitivity; however, as expected
SSSD significantly increased plasma concentration of total
cholesterol and triacylglycerol as compared to NCSD and
plasma triacylglycerol concentration as compared to
WATER. The fact that the study subjects remained weight-
stable throughout the intervention indicated that the results
were independent of changes in body weight. In compar-
ison, Maki et al. found a beneficial effect of dairy in relation
to insulin sensitivity in a 2 × 6 week crossover study
Milk and insulin sensitivity

Table 4 Average daily

Milk SSSD NCSD Water P-value
consumption of energy and
macronutrients at baseline and at Total energy, kJ
the end of the 6 months
intervention and difference Pre-intervention 9931 ± 2549 10,834 ± 2648 9592 ± 2611 10,424 ± 2289 0.58
between the four beverage Post-intervention 10,830 ± 889 9657 ± 890 9495 ± 660 10,638 ± 806 0.14
groupsa Fat, % of energy
Pre-intervention 33.1 ± 5.9 32.7 ± 3.5 32.4 ± 3.4 34.6 ± 4.7 0.56
Post-intervention 32.8 ± 1.2b 31.3 ± 1.0b 34.2 ± 1.4b,c 38.5 ± 1.0c <0.01
Total fat, g
Pre-intervention 89.4 ± 30.0 95.0 ± 24.6 84.4 ± 26.9 99.0 ± 31.9 0.53
Post-intervention 96.7 ± 10.0 81.3 ± 7.5 87.6 ± 6.8 112.5 ± 11.3 0.09
Saturated fat
Pre-intervention 32.1 ± 12.3 31.1 ± 9.6 27.5 ± 10.2 32.8 ± 13.7 0.61
Post-intervention 37.8 ± 4.1 24.9 ± 2.2 28.7 ± 2.4 36.7 ± 4.3 0.05
Monounsaturated fat
Pre-intervention 26.8 ± 9.0 26.7 ± 7.8 24.8 ± 10.7 27.7 ± 10.2 0.87
Post-intervention 27.1 ± 2.8 22.0 ± 1.9 24.8 ± 2.0 32.9 ± 3.9 0.06
Polyunsaturated fat
Pre-intervention 10.8 ± 3.1 12.4 ± 2.9 12.2 ± 4.7 13.3 ± 3.5 0.36
Post-intervention 12.2 ± 1.2 11.4 ± 1.0 13.4 ± 1.1 16.8 ± 1.8 0.13
Carbohydrate, % of energy
Pre-intervention 48.2 ± 6.1 52.1 ± 5.8 48.2 ± 4.5 48.5 ± 4.8 0.20
Post-intervention 47.0 ± 1.0c 53.8 ± 1.9b 46.4 ± 1.1c 44.0 ± 1.3c <0.001
Protein, % of energy
Pre-intervention 16.7 ± 3.2 14.1 ± 2.6 16.2 ± 2.5 14.8 ± 2.6 0.06
Post-intervention 17.9 ± 0.6 c
12.9 ± 0.8 b
15.3 ± 0.6 b,d
15.6 ± 0.6 c,d
<0.001
Calcium, mg
Pre-intervention 1355 ± 436b 958 ± 320c 983 ± 384c 950 ± 290c 0.01
Post-intervention 1848 ± 112 b
671 ± 91 c
1021 ± 128 c
845 ± 62c 0.000
Alcohol, g
Pre-intervention 6.9 ± 8.3 4.3 ± 6.6 11.6 ± 15.9 8.3 ± 10.0 0.37
Post-intervention 9.7 ± 2.8 7.7 ± 4.0 12.4 ± 2.9 7.7 ± 2.2 0.82
NCSD non-caloric soft drink, SSSD sugar-sweetened soft drink
a
All values are means ± SEs. Statistical differences were analyzed in an ANCOVA model with baseline diet
registration from before intervention included as covariates. Values that have no superscript in common are
significantly different from each other (Tukey’s HSD, P < 0.05). n = 53 (two subjects had no dietary
records done and five had nonsufficient records). Data were assessed with a 7-day weighted dietary record
estimated by using Dankost Pro dietary assessment software (Dankost)

comparing dairy (474 mL/d 2% milk and 170 g/d low-fat
yoghurt) with sugar-sweetened products (710 mL/d non-diet
soda and 108 g/d non-dairy pudding) in 33 adults at risk of
T2D. The study found less favorable values for insulin
sensitivity (HOMA) and a liquid meal tolerance test during
the sugar-sweetened product period compared to dairy [19].
Although, the amount of sugar-sweetened beverage intake
was similar to the present study, the dairy was a mix of milk
and yoghurt, and it was therefore not possible to differ-
entiate the effect. Recent large meta-analysis of subtypes of
dairy products reported no association between milk intake
and risk of T2D [4, 5] and a Mendelian randomization study
also confirmed these results [20]. Overall, current evidence
therefore does not confirm a specific effect of milk on
insulin sensitivity in healthy overweight and obese subjects.
Evidence from observational studies support a strong
association between intake of SSSD and risk of T2D,
independent of adiposity [12]. In the primary paper, results
showed that SSSD increased ectopic fat storage compared
to the other beverages [14]. Ectopic fat accumulation by
increased fructose metabolism in the liver is believed to be
one biological mechanism linking SSSD to increased risk of
T2D [21]; however, this did not reflect in risk markers of
T2D after 6-month SSSD intake.

Results showed no difference between MILK and SSSD
in blood lipids. This is in line with the study by Maki et al.
[19], who found no difference in lipid concentration
between the diets, except a decrease (4.2%) in HDL cho-
lesterol during the sugar-sweetened product diet compared
to dairy. SSSD increased plasma triacylglycerol con-
centrations compared to NCSD and WATER, and total
cholesterol concentrations compared to NCSD. This implies
that MILK is beneficial compared to SSSD regarding risk of
CVD. Overall, the unbeneficial effects of SSSD on CVD is
well-established [22–24]. In the present study, we showed
no difference between the beverage groups in blood pres-
sure. RCTs have shown a beneficial effect of milk intake on
blood pressure; however other studies along with the pre-
sent study found no decrease in blood pressure [25, 26].
The dietary records showed that the mean difference in
SFAs corresponded to the difference expected (~10 g/d)
based on nutritional content of the beverages for SSSD and
NCSD but not for WATER. The lack of difference between
WATER and MILK in SFAs could be explained by an
overall high-fat intake (38.5%) with WATER compared to
the other beverage groups.
The strengths of the present RCT were the long duration
including the thorough estimation of risk of T2D with an
OGTT. It is, however, a limitation that as a secondary
analysis, no power calculation was performed for the
OGTT. The free-living design of the study was a limitation
due to potential confounding by other dietary and lifestyle
factors. Moreover, the study was not blinded because the
appearance of the test beverages could not be concealed.
In conclusion, the results of the present study showed no
difference between 6 months' intake of MILK, SSSD,
NCSD, and WATER in effect on risk markers of T2D in
overweight and obese adults. As a secondary analysis, these
results need confirmation in future studies.
Acknowledgements We thank Christian Ritz for support with the
statistical analyses. We gratefully acknowledge the work of Maria
Maersk and Anita Belza, who conducted the intervention and helped
planning the study. The authors’ responsibilities were following—S.E.:
performed the statistical analysis, wrote the manuscript with primary
responsibility for final content; J.M.B.: responsible for the analysis of
PAI-1 and FFA; T.T. and A.R.: initiated the current analyses and
supplied valuable knowledge and scientific consultation during the
process; A.A. and B.R.: took part in the funding and design; all
authors: read and approved the final manuscript. A.A. has received
research grants from Arla Foods, Denmark; The Danish Dairy
Research Foundation; Global Dairy Platform, USA; and the Danish
Agriculture and Food Foundation. T.T. has received research grants
from Arla Foods, Denmark; The Danish Dairy Research Foundation;
and the Dairy Institute, Rosemont, IL. A.R. has received research
funding from the Dairy Research Industry, Rosemont, IL, and The
Danish Agriculture and Food Council.
Compliance with ethical standards
Conflict of interest The authors declare no conflict of interest.
S. Engel et al.
Disclaimer The sponsors had no influence on the execution of the
study, the manuscripts, or its conclusions.
References
1 WHO. WHO Global report on Diabetes. World Health Organiza-
tion. 2016. Internet: Available from: http://apps.who.int/iris/
bitstream/10665/204871/1/9789241565257_eng.pdf (accessed 26
September 2017).
2 Qin L-Q, Xu J-Y, Han S-F, Zhang Z-L, Zhao Y-Y, Szeto IMY.
Dairy consumption and risk of cardiovascular disease: an updated
meta-analysis of prospective cohort studies. Asia Pac J Clin Nutr.
2015;24:90–100.
3 Soedamah-Muthu SS, Ding EL, Al-Delaimy WK, Hu FB,
Engberink MF, Willett WC, et al. Milk and dairy consumption and
incidence of cardiovascular diseases and all-cause mortality: dose-
response meta-analysis of prospective cohort studies. Am J Clin
Nutr. 2011;93:158–71.
4 Aune D, Norat T, Romundstad P, Vatten LJ. Dairy products and
the risk of type 2 diabetes: a systematic review and dose-response
meta-analysis of cohort studies. Am J Clin Nutr. 2013;98:1066–83.
5 Gijsbers L, Ding EL, Malik VS, de Goede J, Geleijnse JM,
Soedamah-Muthu SS. Consumption of dairy foods and diabetes
incidence: a dose-response meta-analysis of observational studies.
Am J Clin Nutr. 2016;103:1111–24.
6 Chen M, Sun Q, Giovannucci E, Mozaffarian D, Manson JE,
Willett WC, et al. Dairy consumption and risk of type 2 diabetes: 3
cohorts of US adults and an updated meta-analysis. BMC Med.
2014;12:211–15.
7 Fonolla J, Lopez-Huertas E, Machado FJ, Molina D, Alvarez I,
Marmol E, et al. Milk enriched with ‘healthy fatty acids’ improves
cardiovascular risk markers and nutritional status in human
volunteers. Nutrition. 2009;25:408–14.
8 Drouin-Chartier JP, Gagnon J, Labonte ME, Desroches S, Charest
A, Grenier G, et al. Impact of milk consumption on cardiometa-
bolic risk in postmenopausal women with abdominal obesity. Nutr
J. 2015;14:12.
9 Soerensen KV, Thorning TK, Astrup A, Kristensen M, Lorenzen
JK. Effect of dairy calcium from cheese and milk on fecal fat
excretion, blood lipids, and appetite in young men. Am J Clin
Nutr. 2014;99:984–91.
10 Barr SI, McCarron DA, Heaney RP, Dawson-Hughes B, Berga SL,
Stern JS, et al. Effects of increased consumption of fluid milk on
energy and nutrient intake, body weight, and cardiovascular risk
factors in healthy older adults. J Am Diet Assoc.
2000;100:810–17.
11 Greenwood DC, Threapleton DE, Evans CE, Cleghorn CL, Nyk-
jaer C, Woodhead C, et al. Association between sugar-sweetened
and artificially sweetened soft drinks and type 2 diabetes: sys-
tematic review and dose-response meta-analysis of prospective
studies. Br J Nutr. 2014;112:725–34.
12 Malik VS, Popkin BM, Bray GA, Despres JP, Willett WC, Hu FB.
Sugar-sweetened beverages and risk of metabolic syndrome and
type 2 diabetes: a meta-analysis. Diabetes Care. 2010;33:2477–83.
13 Rogers PJ, Hogenkamp PS, de Graaf C, Higgs S, Lluch A, Ness
AR, et al. Does low-energy sweetener consumption affect energy
intake and body weight? A systematic review, including meta-
analyses, of the evidence from human and animal studies. Int J
Obes. 2016;40:381–94.
14 Maersk M, Belza A, Stodkilde-Jorgensen H, Ringgaard S, Cha-
banova E, Thomsen H, et al. Sucrose-sweetened beverages
increase fat storage in the liver, muscle, and visceral fat depot: a 6-
mo randomized intervention study. Am J Clin Nutr.
2012;95:283–89.
Milk and insulin sensitivity
15 Bruun JM, Maersk M, Belza A, Astrup A, Richelsen B. Con-
sumption of sucrose-sweetened soft drinks increases plasma levels
of uric acid in overweight and obese subjects: a 6-month rando-
mised controlled trial. Eur J Clin Nutr. 2015;69:949–53.
16 Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained
from oral glucose tolerance testing: comparison with the eugly-
cemic insulin clamp. Diabetes Care. 1999;22:1462–70.
17 Kaplan W. Background Paper 6.4 Diabetes. Priority Medicines for
Europe and the World. ‘A Public Health Approach to Innovation’.
World health Organization. 2013. Internet: Available from: http://
www.who.int/medicines/areas/priority_medicines/BP6_4DM.pdf
(accessed 26 Sseptember 2017).
18 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF,
Turner RC. Homeostasis model assessment: insulin resistance and
beta-cell function from fasting plasma glucose and insulin con-
centrations in man. Diabetologia. 1985;28:412–19.
19 Maki KC, Nieman KM, Schild AL, Kaden VN, Lawless AL,
Kelley KM, et al. Sugar-sweetened product consumption alters
glucose homeostasis compared with dairy product consumption in
men and women at risk of type 2 diabetes mellitus. J Nutr.
2015;145:459–66.
20 Bergholdt HK, Nordestgaard BG, Ellervik C. Milk intake is not
associated with low risk of diabetes or overweight-obesity: a
Mendelian randomization study in 97,811 Danish individuals. Am
J Clin Nutr. 2015;102:487–96.
21 Malik VS. Sugar sweetened beverages and cardiometabolic health.
Curr Opin Cardiol. 2017;32:572–79. 1
22 Xi B, Huang Y, Reilly KH, Li S, Zheng R, Barrio-Lopez MT, et al.
Sugar-sweetened beverages and risk of hypertension and CVD: a
dose-response meta-analysis. Br J Nutr. 2015;113:709–17.
23 Richelsen B. Sugar-sweetened beverages and cardio-metabolic
disease risks. Curr Opin Clin Nutr Metab Care. 2013;16:478–84.
24 Te Morenga LA, Howatson AJ, Jones RM, Mann J. Dietary sugars
and cardiometabolic risk: systematic review and meta-analyses of
randomized controlled trials of the effects on blood pressure and
lipids. Am J Clin Nutr. 2014;100:65–79.
25 Benatar JR, Sidhu K, Stewart RAH. Effects of high and low fat
dairy food on cardio-metabolic risk factors: a meta-analysis of
randomized studies. PLoS ONE. 2013;8:e76480.
26 Crichton GE, Howe PRC, Buckley JD, Coates AM, Murphy KJ.
Dairy consumption and cardiometabolic health: outcomes of a 12-
month crossover trial. Nutr Metab. 2012;9:19.