Content uploaded by Nicole Neufingerl
Author content
All content in this area was uploaded by Nicole Neufingerl on Oct 26, 2015
Content may be subject to copyright.
Effect of cocoa and theobromine consumption on serum
HDL-cholesterol concentrations: a randomized controlled trial
1–3
Nicole Neufingerl, Yvonne EMP Zebregs, Ewoud AH Schuring, and Elke A Trautwein
ABSTRACT
Background: Evidence from clinical studies has suggested that
cocoa may increase high-density lipoprotein (HDL)–cholesterol
concentrations. However, it is unclear whether this effect is attribut-
able to flavonoids or theobromine, both of which are major cocoa
components.
Objectives: We investigated whether pure theobromine increases
serum HDL cholesterol and whether there is an interaction effect
between theobromine and cocoa.
Design: The study had a 2-center, double-blind, randomized, pla-
cebo-controlled, full factorial parallel design. After a 2-wk run-in
period, 152 healthy men and women (aged 40–70 y) were randomly
allocated to consume one 200-mL drink/d for 4 wk that contained 1)
cocoa, which naturally provided 150 mg theobromine and 325 mg
flavonoids [cocoa intervention (CC)], 2) 850 mg pure theobromine
[theobromine intervention (TB)], 3) cocoa and added theobromine,
which provided 1000 mg theobromine and 325 mg flavonoids [theo-
bromine and cocoa intervention (TB+CC)], or 4) neither cocoa nor
theobromine (placebo). Blood lipids and apolipoproteins were mea-
sured at the start and end of interventions.
Results: In a 2-factor analysis, there was a significant main effect of
the TB (P,0.0001) but not CC (P= 0.1288) on HDL cholesterol
but no significant interaction (P= 0.3735). The TB increased HDL-
cholesterol concentrations by 0.16 mmol/L (P,0.0001). Further-
more, there was a significant main effect of the TB on increasing
apolipoprotein A-I (P,0.0001) and decreasing apolipoprotein B
and LDL-cholesterol concentrations (P,0.02).
Conclusions: Theobromine independently increased serum HDL-
cholesterol concentrations by 0.16 mmol/L. The lack of significant
cocoa and interaction effects suggested that theobromine may be the
main ingredient responsible for the HDL cholesterol–raising effect.
This trial was registered at clinicaltrials.gov as NCT01481389.
Am J Clin Nutr 2013;97:1201–9.
INTRODUCTION
A number of observational and clinical studies indicated that
cocoa or cocoa-containing products may reduce cardiovascular
disease (CVD)
4
risk (1) and improve CVD-related risk factors
such as blood pressure (2), LDL oxidation, inflammatory status
(3, 4), and the blood lipid profile (5–7). Several clinical studies
showed that cocoa increases HDL-cholesterol concentrations
(8–13), although other studies did not confirm such a beneficial
effect (14–20). Low serum HDL cholesterol (,1 mmol/L) is
considered an independent and inverse CVD risk factor, and
beyond lowering LDL cholesterol, increasing HDL cholesterol
has been suggested as a secondary lipid target for reducing CVD
risk (21). The cardioprotective effects of cocoa are commonly
attributed to cocoa flavonoids, which have been reported to in-
fluence various CVD risk factors through multiple mechanistic
pathways (3, 4, 22, 23). However, mechanistic evidence for
cocoa flavonoids to affect blood lipids is lacking (4, 22). In
particular, it is not clear whether flavonoids or possibly another
bioactive component of cocoa (ie, theobromine) are responsible
for the reported increase in serum or plasma HDL-cholesterol
concentrations. Strikingly, clinical studies that compared
theobromine-containing cocoa products to theobromine-free
control products repeatedly showed significant beneficial effects
on HDL cholesterol (8–12). In contrast, studies with test prod-
ucts that controlled for the theobromine content mostly did not
show an effect on increasing HDL cholesterol (14, 16–18, 20).
These results suggested that theobromine in cocoa, rather than
flavonoids, may be involved or even responsible for the HDL-
cholesterol increasing benefit.
Therefore, to investigate whether theobromine by itself or
together with cocoa can increase HDL-cholesterol concentra-
tions, we studied the effects of cocoa, pure theobromine, and the
combination of cocoa with added theobromine on changes in the
blood lipid profile in healthy humans. Because theobromine is
known for its vasodilating and cardiac-stimulating function (24,
25), we also investigated possible effects of theobromine on
blood pressure and heart rate.
SUBJECTS AND METHODS
Subjects
Subjects within the subject database of Eurofins Optimed
Clinical Research, Grenoble and Lyon, France, who met the
major inclusion criteria were contacted and invited for a screening
1
From the Nutrition and Health Department (NN and EAT) and Unilever
Clinicals (YEMPZ and EAHS), Unilever Research & Development Vlaar-
dingen, Vlaardingen, Netherlands.
2
This study was funded by Unilever Research & Development Vlaardingen.
3
Addresscorrespondence to N Neufingerl, Olivier van Noortlaan 120, 3130
AC Vlaardingen, Netherlands. E-mail: nicole.neufingerl@unilever.com.
4
Abbreviations used: ABPM, ambulatory blood pressure monitor; CC,
cocoa intervention; CETP, cholesteryl ester transfer protein; CVD, cardio-
vascular disease; FAS, full-analysis-set; PP, per protocol; TB, theobromine
intervention; TB+CC, theobromine and cocoa intervention.
Received July 17, 2012. Accepted for publication March 15, 2013.
First published online April 17, 2013; doi: 10.3945/ajcn.112.047373.
Am J Clin Nutr 2013;97:1201–9. Printed in USA. Ó2013 American Society for Nutrition 1201
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
visit. In total, 152 healthy men and postmenopausal women, aged
40–70 y, participated in the study. Participants had 10-y risk of
developing CVD ,10% based on the European low-risk chart
(26); their blood lipids, blood pressure, heart rate, blood glucose,
and other biochemical and hematologic variables were within
healthy reference ranges as judged by the research physician.
Subjects with diabetes or previous cardiovascular events as well
as smokers were excluded from the study. Also, subjects with
a reported habitual consumption .8 caffeine-containing drinks/d day,
high alcohol consumption (.14 units/wk for women; .21 units/wk
for men), or intense sporting activities (.10 h/wk) were ex-
cluded. Study participants were informed about the aim of the
study, and written informed consent was obtained from each
participant before the start of the study. The study was approved
and conducted in accordance with the guidelines of the ethics
committee Comite
´de Protection des Personnes Sud-Est IV,
Lyon, France.
Study design
The study had a double-blind, randomized, placebo-controlled,
full factorial parallel design with a 2-wk run-in period followed
by an intervention period of 4 wk. It was a bicentric study that was
carried out at 2 sites in Grenoble and Lyon, France, between
December 2010 and February 2011. Participants who passed the
screening and completed the run-in period were sequentially
allocated by the clinical investigator into 4 treatment groups
according to a preestablished blockwise randomization scheme.
Separate computer-generated randomization schedules for men
and women were created by a Unilever statistician to stratify
subjects by sex with an attempt to achieve a distribution between
men and women of 50:50 in each treatment group.
Study products
During the intervention period, participants were instructed to
consume one of the following test drinks daily for 4 wk: a 200-mL
drink that contained 1) 6 g cocoa powder that naturally provided
150 mg theobromine and 325 mg flavonoids ([cocoa in-
tervention (CC)], 2) 850 mg pure theobromine [theobromine
intervention (TB)], 3) 6 g cocoa powder and 850 mg added pure
theobromine, which provided a total of 1000 mg theobromine
and 325 mg flavonoids [theobromine and cocoa intervention (TB
+CC)], or 4) no cocoa powder and no added theobromine
(placebo). Test drinks were chocolate-flavored pasteurized
acidified milk drinks that were manufactured by Unilever Re-
search & Development. The various test drinks differed to some
extent in taste and color. To maintain the blinding of participants
and investigators, drinks were supplied in identical tinted bottles
that were packed individually for each participant in a neutral
box and labeled with the subject code; also participants were
instructed not to pour the drink into a glass but to consume it
directly out of the tinted bottle. All test drinks provided w80
kcal, 3 g fat, 11 g carbohydrates, and 3 g protein per 200-mL
drink. Theobromine was supplied by Fagron; the cocoa powder
(Acticoa) was provided by Barry Callebaut. The theobromine,
total polyphenol, and cocoa flavonoid (ie, flavanol) concentra-
tion of the cocoa powder and test drinks were determined before
the start and at the end of the study. Total polyphenols were
analyzed by using the Folin-Ciocalteau method (27), which was
adapted to microplate format. Theobromine and flavanol con-
tents were quantified with the HPLC method according to In-
ternational Standardization Organization standard 14502–2:2005
(E). The response factor for theobromine was determined for
correct quantification.
Study procedures
Subjects were instructed to consume the test drink 1 h before
the consumed breakfast. Subjects were asked to shake the drink
$20 times before use to ensure that the cocoa powder was well
dissolved before ingestion.
Throughout both the run-in and intervention periods, subjects
were requested to exclude all cocoa-containing products (such
as, eg, chocolates, chocolate bars, chocolate drinks, chocolate
cookies, chocolate spread, chocolate ice cream, chocolate pud-
ding, chocolate cake, and breakfast cereals with chocolate) from
their diets. The consumption of caffeine-containing drinks such
as coffee, tea, or cola was restricted to a maximum of 4 cups or
cans/d because in the body, caffeine is metabolized into theo-
bromine. The use of supplements that contained theobromine,
caffeine, or cocoa and consumption of caffeine-containing energy
drinks was prohibited. Beyond these limitations, subjects were
asked to maintain their habitual diet and lifestyle routines.
Subjects were instructed to report daily in an intake diary the
amount of test product taken, timings of test product and
breakfast intakes, and compliance to restrictions concerning the
background diet. The importance of the compliance to these
instructions was also emphasized during weekly phone calls to
study participants. At the end of the intervention period, subjects
were asked to return all (used and nonused) bottles of the test
product to the study center.
Subjects visited the study center on 2 consecutive days at
baseline (ie, on the last day of the run-in period and the first
day of the intervention period) and again on 2 consecutive days
at the end of the intervention period (ie, on the last day on which
the test product was consumed and the day after). On all 4 mea-
surement days, subjects came to the study center in a fasted state
not having eaten anything since 2000 and not having drunk
anything (except water) after 2200 the previous evening. Breakfast
was provided to subjects at the study center 1 h after intake of the
test product.
Clinical and laboratory measurements
For the measurement of blood lipids, blood samples were taken
on each of the 4 measurement days to take into account the day-
to-day variation in blood lipids. Samples taken on the first day of
the intervention and the day after the end of the intervention
period were used to determine plasma theobromine and caffeine
concentrations as additional compliance measures. Measurement
days were not scheduled on Mondays to avoid effects of possible
deviating behaviors during weekends on blood measurements.
With the use of a venipuncture, fasted blood samples for the
measurement of blood lipids were collected into SST II advance
BD tubes (Becton Dickinson), and aliquots were transferred into
prelabeled polypropylene tubes and frozen at 2208C. At the end
of the study, sample tubes were sent to the Laboratoire Biocentre
for analysis of blood lipids and lipoproteins. Serum HDL cho-
lesterol, LDL cholesterol, total cholesterol, triacylglycerides,
1202 NEUFINGERL ET AL
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
apolipoprotein A-I, and apolipoprotein B were measured with
a Hitachi 912 autoanalyzer with the aid of test kits, all provided
by Roche Diagnostics GmbH.
For the analysis of theobromine and caffeine plasma con-
centrations, fasted blood samples were collected into EDTA-3K–
coated evacuated tubes and centrifuged at 1500 3gfor 10 min
at 48C. The top layer of human plasma was transferred into 3
prelabeled polypropylene tubes and frozen at 2808C. Theo-
bromine and caffeine concentrations were analyzed at Unilever
Research & Development. To 400 mL plasma samples, 50 mL
internal standard (theobromine-d
6
) and 750 mL acetonitrile were
added. After mixing and centrifugation, the supernatant fluid
was analyzed by using liquid chromatography–tandem mass
spectrometry (28). All serum and plasma samples of one subject
were analyzed in the same batch.
Body weight was measured at baseline (ie, on the first day of
the intervention) and the day after the end of the intervention
period early in the morning after an overnight fast. Subjects were
weighed in their underwear and with an empty bladder.
In a subsample of the study population (n= 10 per treatment
group), 24-h ambulatory blood pressure and heart rate were
measured on the last day of the run-in period and the last day of
the intervention. Measurements were performed by using an
automatic ambulatory blood pressure monitor (ABPM) (Space-
labs 90207; Spacelabs Inc), which was placed on the non-
dominant arm of subjects. The ABPM device measured systolic
blood pressure, diastolic blood pressure, and heart rate every 15
min during the day (0700–2300) and every 30 min during the
night (2300–0700). The first ABPM measurement set off in the
morning after blood sampling and measurements continued for
24 h. The 40 subjects who underwent the 24-h blood pressure
and heart rate measurements were instructed not to consume
alcohol-containing beverages, tea, coffee, or grape juice or
conduct strenuous physical activities during measurement days
to reduce external influences that may have affected blood
pressure and heart rate.
Adverse events were monitored throughout the study period,
ie, they were spontaneously reported by the subjects or noted by
the investigator during one of the visits to the study center.
Statistical analysis
A power calculation indicated that a minimum of 138 subjects
were necessary to detect an interaction effect on HDL-cholesterol
concentrations of 0.07 mmol/L with an SD of 0.15 in a 2-by-2 full
factorial design, testing 2 sided (ie, for a positive or negative
effect) with a= 0.05 and b= 0.8. To allow for a dropout rate of
10%, the desired sample size was 152.
Mean values for blood lipid concentrations at baseline and at
the end of the intervention were calculated as the average of the
measurements taken on the 2 consecutive measurement days.
Values for non-HDL cholesterol were calculated by subtracting
HDL cholesterol from total cholesterol. HDL-cholesterol:non–
HDL-cholesterol and HDL-cholesterol:LDL-cholesterol ratios
were determined. Mean 24-h values of diastolic and systolic
blood pressure and heart rate were calculated by first taking
hourly means of blood pressure or heart rate measurements and
discarding any missing values. Hourly means were averaged
over 24 h. If $30% of hourly means were missing, no mean
24-h value was calculated, and values were reported as missing
completely. Before the analysis, all data were checked for
being normally distributed. Triacylglycerol concentrations and
HDL-cholesterol:non–HDL-cholesterol and HDL-cholesterol:
LDL-cholesterol ratios, and plasma theobromine and caffeine
concentrations required log transformation.
Blood lipid and apolipoprotein concentrations and mean 24-h
systolic and diastolic blood pressure and heart rate were analyzed
for factor effects by using a full factorial analysis with baseline
and sex as covariates and factors split up in cocoa (yes or no) and
theobromine (yes or no). Also, sex-by-treatment interactions
were included in the base model. In addition, age, body weight,
and BMI at baseline and changes in body weight and BMI were
tested as potential covariates. Variables were only to be included
in the model if they were shown to significantly add to the
explained variance as measured by using the chi-square-test in
the ANOVA table (P#0.1). Changes in compliance markers (ie,
theobromine and caffeine plasma concentrations and body
weight) from baseline to the end of the intervention within and
between groups were analyzed by using paired sample ttests
and ANOVA with Tukey’s multiple-difference testing, re-
spectively. P,0.05 was considered statistically significant.
In addition to a full-analysis-set (FAS) analysis that included
all subjects for which end-of-intervention values were available,
a per-protocol (PP) analysis was performed, which excluded data
from subjects who had been noncompliant with test-product
consumption (defined as ,80% of the total intake of test
products), had a body weight change .2 kg, or were noncom-
pliant with dietary restrictions (ie, consumed cocoa-containing
products in unreasonably high amounts or toward the end of the
intervention). These exclusion criteria were applied because
noncompliance with test-product intake and the background diet
may have blunted the effects of the TB, and weight changes can
influence HDL-cholesterol concentrations (29). All statistical
analyses were performed with JMP software (version 9, SAS
Institute Inc).
RESULTS
Subject characteristics and compliance measures
Of the 152 subjects randomly assigned over the 4 treatments,
143 subjects completed the intervention period and were avail-
able for FAS analysis. This analysis excluded 9 subjects who had
discontinued the intervention because of the occurrence of ad-
verse events (n= 8) or personal reasons (n= 1). A flow diagram
of participants throughout the study is depicted in Figure 1.
During a blind review, 9 additional subjects were excluded from
the PP analyses because they had gained or lost .2 kg in body
weight during the study period. Because the PP and FAS anal-
yses showed similar results, only the results of the FAS pop-
ulation (n= 143) are reported.
Baseline characteristics of study participants and key com-
pliance measures per treatment group are shown in Table 1. Men
and women were evenly distributed across and within treatment
groups; baseline characteristics were similar between subjects of
different groups. The compliance to test-product intake as cal-
culated from the counting of returned empty bottles and re-
cordings in intake diaries was 99.7%. From baseline to the end
of the intervention, plasma concentrations of theobromine in-
creased 5-, 17- and 19-fold in the CC, TB, and TB+CC groups,
THEOBROMINE RAISES HDL CHOLESTEROL 1203
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
respectively. Plasma caffeine concentrations at baseline and at
the end of the intervention were generally low in all groups (ie,
,4mmol/L), although during the intervention period, caffeine
concentrations doubled in the TB+CC group (Table 1).
Within but not between groups, there were small, significant
changes in the mean body weight from baseline to the end of the
intervention, which was, however, not considered to be of bi-
ological relevance (Table 1). The body weight, BMI, or age of
FIGURE 1. Subject flowchart. Data from all subjects for whom baseline and follow-up measurements were available (ie, FAS) were included in the
analysis. *Numbers in italic represent the study population for blood pressure and heart rate measurements. FAS, full analysis set; PP, per protocol.
TABLE 1
Baseline characteristics of study participants and key compliance measures
Placebo Cocoa Theobromine Theobromine plus cocoa
n
1
38 38 38 38
Men (%) 50 50 50 50
Age (y) 55.4 68.7 (40–70)
2
55.2 68.2 (40–67) 53.3 69.4 (40–69) 55.9 68.0 (40–70)
BMI (kg/m
2
) 24.9 63.1 (19.6–30.8) 24.8 62.9 (19.1–30.0) 23.8 62.5 (18.8–28.4) 24.3 62.7 (18.9–29.4)
Body weight (kg)
Baseline 72.2 613.2 (53.9–100.7) 70.5 612.4 (47.6–91.4) 67.9 610.7 (46.0–90.5) 69.2 611.9 (48.9–93.1)
Change (4 wk –
baseline)
0.3 61.1 (23.0 to 3.9) 0.8 61.0 (21.6 to 4.1)
3
0.5 61.1 (22.3 to 2.8)
4
0.6 61.0 (22.4 to 3.0)
4
Plasma theobromine
(mmol/L)
Baseline 1.85 61.87 (,0.5 to 10.00) 1.76 61.22 (,0.5 to 4.90) 1.77 61.78 (,0.5 to 10.30) 2.01 62.35 (,0.5 to 11.50)
Change (4 wk –
baseline)
0.30 62.65 (25.60 to 10.45) 7.17 66.42 (1.20–38.40)
3
28.75 616.12 (8.50–77.45)
3,5
37.12 621.39 (6.20–92.00)
3,5
Plasma caffeine (mmol/L)
Baseline 2.13 64.88 (,0.5 to 28.40) 2.46 63.51 (,0.5 to 18.80) 1.79 62.15 (,0.5 to 7.3) 1.78 62.74 (,0.5 to 10.80)
Change (4 wk –
baseline)
20.81 64.84 (227.4 to 5.70) 0.29 62.54 (23.55 to 10.70) 0.74 62.68 (25.75 to 10.90) 1.83 63.87 (27.2 to 14)
4,6
1
Total number of subjects was n= 152 at baseline and n= 143 at the end of the intervention. There were no significant differences between groups at
baseline (P.0.05; ANOVA).
2
Mean 6SD; ranges in parentheses (all such values).
3,4
Significantly different from baseline (paired samples ttest):
3
P,0.0001,
4
P,0.05.
5
Significantly different from cocoa and placebo groups, P,0.0001 (ANOVA followed by Tukey’s multiple-difference testing).
6
Significantly different from the placebo group, P,0.05 (ANOVA followed by Tukey’s multiple-difference testing).
1204 NEUFINGERL ET AL
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
subjects were NS predictors of any outcome measures; therefore,
only sex and the baseline value of outcome measures were in-
cluded as covariates in the ANCOVA model.
Serum lipids and lipoproteins
Descriptive data on serum lipids and lipoproteins at baseline
and at the end of the intervention for the 4 groups are shown in
Table 2. After 4 wk of intervention, a significant main effect on
HDL-cholesterol concentrations was shown for the TB factor
(P,0.0001) but not for the CC factor (Table 3). The TB in-
creased HDL cholesterol by 0.16 mmol/L. There was no sig-
nificant interaction effect between TB and CC factors on HDL
cholesterol (P= 0.3735). The effect of the TB factor on HDL
cholesterol was similar for men and women.
Significant main effects of the TB but not CC factor were also
shown for LDL cholesterol and the calculated HDL-cholesterol:
non–HDL-cholesterol and HDL-cholesterol:LDL-cholesterol
ratios. There were no significant interaction effects of TB and
CC factors on any of these variables (Table 3). In addition, the
TB factor had a significant main effect on increasing apolipo-
protein A-I concentrations and decreasing apolipoprotein B
concentrations. Also, the CC factor had a small significant effect
on apolipoprotein A-I (Table 3). There were no significant in-
teraction effects between TB and CC factors on apolipoprotein
A-I and apolipoprotein B. No significant factor effects were
shown for total cholesterol and triacylglycerol concentrations
(Table 3).
Twenty-four hour blood pressure, heart rate, and adverse
effects
Descriptive data of 24-h blood pressure and heart rate at
baseline and the end of the intervention for the 4 groups are
shown in Table 4. No significant factor effects were observed for
either 24-h systolic or diastolic blood pressure or 24-h heart rate
(Table 3). However, plotting of the hourly mean data showed
that heart rate acutely increased by 10–15 beats/min during the
first few hours after consumption of the test product in TB and
TB+CC groups (data not shown).
Reported adverse effects were mainly seen in TB and TB+CC
groups, with 43 and 57 recordings, respectively, compared with
18 and 10 recordings in the CC and placebo groups, respectively.
The most commonly reported adverse events were nausea,
vomiting, headache, and diarrhea. All except for one of these
adverse events were of mild to moderate intensity and resolved
before the end of the study. In total, adverse events lead to the
premature withdrawal of 8 subjects, of which 6 subjects were in
the TB+CC group.
DISCUSSION
This randomized, double-blind, placebo-controlled human-
intervention study showed that daily doses of 850 mg theobromine
independently increased serum HDL-cholesterol concentrations by
0.16 mmol/L. Cocoa, which delivered 150 mg natural theobromine
per daily serving, led to only a modest nonsignificant increase in
HDL cholesterol; there was no significant interaction effect be-
tween TB and CC factors on raising HDL cholesterol. These
findings suggested that theobromine is the major active component
in cocoa that is responsible for an increasing HDL-cholesterol
effect. The main effect of the TB factor on increasing HDL
cholesterol was supported by a significant main effect of the TB
factor on an increasing concentration of apolipoprotein A-I, which
is the major apolipoprotein of HDL particles.
The findings of this study might explain conflicting results
reported by previous intervention studies with cocoa on the effect
on HDL-cholesterol concentrations as described in 3 meta-
analyses (5–7). Whereas 2 meta-analyses reported either a non-
significant small reduction (7) or nonsignificant small increase
in HDL-cholesterol concentrations (5) after the intake of cocoa
TABLE 2
Descriptive values of serum lipid and apolipoprotein concentrations at baseline and end of intervention
1
Placebo Cocoa Theobromine Theobromine plus cocoa
n37 37 37 32
HDL cholesterol (mmol/L)
Baseline 1.42 60.38 1.52 60.47 1.45 60.45 1.48 60.43
4 wk 1.56 60.42 1.71 60.51 1.74 60.49 1.86 60.59
LDL cholesterol (mmol/L)
Baseline 3.58 60.88 3.65 60.82 3.21 60.79 3.17 60.80
4 wk 3.62 60.86 3.65 60.73 3.14 60.80 3.07 60.83
Total cholesterol (mmol/L)
Baseline 5.95 61.03 6.09 61.01 5.54 61.00 5.49 60.98
4 wk 5.99 61.05 6.11 60.93 5.65 60.99 5.60 60.85
Triacylglycerol (mmol/L)
Baseline 1.32 60.94 1.16 60.78 1.16 60.48 1.04 60.54
4 wk 1.27 60.88 1.16 60.76 1.22 60.70 0.96 60.47
Apolipoprotein A-I (g/L)
Baseline 1.49 60.27 1.53 60.28 1.48 60.30 1.49 60.27
4 wk 1.44 60.25 1.53 60.26 1.55 60.30 1.55 6027
Apolipoprotein B (g/L)
Baseline 0.87 60.21 0.88 60.20 0.78 60.20 0.77 60.19
4 wk 0.90 60.20 0.89 60.18 0.77 60.20 0.75 60.20
1
All values are means 6SDs. Data from all subjects for whom baseline and follow-up measurements were available
were included.
THEOBROMINE RAISES HDL CHOLESTEROL 1205
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
products, a third meta-analysis reported a significant beneficial
effect of cocoa products on raising HDL cholesterol but with
significant heterogeneity in individual studies (6). A common
feature of studies that showed an HDL-cholesterol–increasing
effect of cocoa was the use of theobromine-free control products
(8–12), whereas studies that controlled for theobromine intake
in control and intervention products did not find an effect (14,
16–18, 20). However, there were a few studies that failed to
show an HDL-cholesterol increasing effect of cocoa despite the
use of a theobromine-free control product (19, 30–33). However,
this lack of an effect may be explained by a too short in-
tervention period (ie, 2 wk) in these studies (30–33) or a too low
theobromine content of the test product (ie, 26 mg per daily
dose) (19). Also, in the current study, the CC treatment, which
naturally contained 150 mg theobromine, caused an only small,
nonsignificant increase in HDL cholesterol, which indicated that
a higher dose of theobromine may be necessary to show a sub-
stantial effect on serum HDL cholesterol.
The prevalence of low serum HDL-cholesterol concentrations
is w26–80% in adult populations across the globe (34). How-
ever, the current study showed that theobromine could increase
HDL cholesterol even in healthy people with adequate basal
TABLE 3
Main factor and interaction effects of theobromine and cocoa
1
Outcome measures and factors Values P
HDL cholesterol (mmol/L)
CC 0.06 60.04 (20.02, 0.14) 0.1288
TB 0.16 60.04 (0.08, 0.24) ,0.0001
TB*CC 0.05 60.06 (20.06, 0.16) 0.3735
LDL cholesterol (mmol/L)
CC 20.02 60.07 (20.17, 0.12) 0.7602
TB 20.18 60.08 (20.33, 20.04) 0.0155
TB*CC 20.00 60.11 (20.21, 0.21) 0.9992
Total cholesterol (mmol/L)
CC 0.03 60.10 (20.16, 0.22) 0.7729
TB 20.04 60.10 (20.23, 0.15) 0.6989
TB*CC 20.04 60.14 (20.31, 0.24) 0.7972
Triacylglycerol (mmol/L)
CC 20.01 60.05 (20.11, 0.08) 0.7893
TB 0.02 60.05 (20.08, 0.12) 0.7243
TB*CC 20.10 60.07 (20.24, 0.04) 0.1805
Apolipoprotein A-I (g/L)
CC 0.06 60.02 (0.01, 0.11) 0.0148
TB 0.11 60.02 (0.06, 0.16) ,0.0001
TB*CC 20.06 60.04 (20.13, 0.01) 0.1133
Apolipoprotein B (g/L)
CC 20.01 60.02 (20.04, 0.02) 0.4466
TB 20.05 60.02 (20.08, 20.02) 0.0015
TB*CC 0.00 60.02 (20.04, 0.05) 0.8681
HDL-cholesterol:non–HDL-cholesterol ratio
CC 0.04 60.03 (20.02, 0.10) 0.1652
TB 0.13 60.03 (0.07, 0.18) ,0.0001
TB*CC 0.03 60.04 (20.05, 0.11) 0.4249
HDL-cholesterol:LDL-cholesterol ratio
CC 0.03 60.03 (20.02, 0.08) 0.1974
TB 0.12 60.03 (0.07, 0.17) ,0.0001
TB*CC 0.02 60.04 (20.06, 0.09) 0.6642
24-h SBP (mm Hg)
CC 20.95 62.71 (26.53, 4.63) 0.7290
TB 2.55 62.55 (22.69, 7.80) 0.3263
TB*CC 1.51 63.91 (26.52, 9.55) 0.7016
24-h DBP (mm Hg)
CC 20.84 61.39 (23.69, 2.01) 0.5486
TB 20.20 61.35 (22.97, 2.57) 0.8821
TB*CC 2.14 62.03 (22.04, 6.32) 0.3025
24-h heart rate (beats/min)
CC 0.80 62.48 (24.27, 5.87) 0.7487
TB 1.99 62.48 (23.07, 7.05) 0.4280
TB*CC 2.77 63.57 (24.52, 10.07) 0.4433
1
All values are mean 6SE estimates; 95% CIs in parentheses. A full factorial analysis with cocoa and theobromine as
full factorial variables and sex and baseline values as covariates was used to test for factor and interaction effects. Data from
all subjects for whom baseline and follow-up measurements were available were included (n= 143 for lipid and apolipo-
protein measures; n= 39 for blood pressure and heart rate measurements). CC, cocoa intervention; DBP, diastolic blood
pressure; SBP, systolic blood pressure; TB, theobromine intervention.
1206 NEUFINGERL ET AL
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
HDL-cholesterol concentrations. The size of the observed effect
of theobromine falls well within the range of effects of raising
HDL cholesterol reported on dietary and lifestyle changes [eg,
regular physical activity (35, 36), modest alcohol consumption
(37), or loss of w5–10% of body weight (29)].
In addition to an increase in HDL cholesterol, supplementation
with theobromine also resulted in a moderate reduction of LDL-
cholesterol concentrations, which further improved the blood
lipid profile. In this study, no significant factor effects on 24-h
blood pressure and heart rate were shown, although an acute
postintake increase in heart rate was observed in TB and TB+CC
groups. This cardiac stimulating effect, which is an undesirable
side effect, was also reported in previous studies with theobro-
mine or other methylxanthines (24, 25).
A strength of our study was the high quality of execution and
data generation; ,6% of participants withdrew from the study
prematurely. In addition, the reported compliance to test-product
intake was very high and was further supported by observed
increases in theobromine plasma concentrations in active treat-
ment groups. The doubling of plasma caffeine concentrations in
the TB+CC group may be explained by the competition between
caffeine and theobromine, which are both methylxanthines, for
metabolic pathways.
Because HDL cholesterol can be influenced by diet and
lifestyle behaviors, such as physical exercise and alcohol con-
sumption, we excluded subjects with extreme physical activity
or high alcohol intake from participation; through daily self-
recordings of compliance, we discouraged changes in and captured
deviations from the habitual diets and lifestyles of participants
throughout the study period. However, the lack of information on
actual physical activity and a complete dietary assessment can be
seen as a limitation of our study because the diets and lifestyles of
subjects in the 4 treatment groups may have differed at baseline or
changed differentially throughout the study period.
Another limitation of this study is that we did not investigate
possible physiologic mechanisms of action that could explain
how theobromine increases HDL cholesterol. HDL-cholesterol
metabolism is rather complex, and several biological targets are
known to affect serum HDL-cholesterol concentrations, such as
the enhancement of apolipoprotein A-I production; enhancement
of ATP-binding cassette transporter A1/G1 activity, or inhibition
of cholesteryl ester transfer protein (CETP) activity. Post-hoc, we
did analyze CETP activity in remaining plasma samples of
subjects in the placebo and TB group, showing no effect of
theobromine on CETP activity (data not shown). Clearly, addi-
tional studies are required to address the effects of theobromine
on possible biological targets to explore the underlying mech-
anisms of action for the increase in HDL cholesterol. In addi-
tion, there is no current indication of whether the increase in
HDL-cholesterol concentrations with theobromine affects HDL
functionality.
A number of epidemiologic studies, including prospective
observation studies, have shown that a 1% increase in HDL
cholesterol is linked to a 1–3% reduction in CVD risk, even
when controlled for other risk variables (38–40). Although el-
evated LDL cholesterol is the most important lipid target in the
prevention and treatment of CVD, and despite powerful treat-
ment options to lower LDL-cholesterol concentrations, there
remains substantial residual cardiovascular risk, which can be
further reduced by targeting other blood lipids such as raising
HDL cholesterol (41). In addition, epidemiologic data have shown
that increasing HDL-cholesterol concentrations can further re-
duce CVD risk even in people with adequate LDL-cholesterol
concentrations (42, 43). However, there is currently no direct
clinical evidence that increasing HDL-cholesterol concentra-
tions per se has a protective effect on reducing CVD risk.
Still, the observed beneficial effects of theobromine on blood
lipids suggested that theobromine may be a promising dietary
ingredient in the prevention of CVD. Chocolate is by far the main
source of theobromine in the Western diet (25). However, to
deliver 850 mg theobromine (ie, the daily dose which was
provided by the TB treatment in the current study), w100 g dark
chocolate or w200 g milk chocolate are needed (25). Never-
theless, this amount of chocolate also provides w20–40 g sat-
urated fat and w540–1,080 kcal, respectively (44). Therefore,
other ways to increase dietary theobromine intake are required.
In conclusion, to our knowledge, this is the first study to show
that a daily intake of 850 mg theobromine independently and
significantly increases HDL-cholesterol concentrations by 0.16
mmol/L in healthy subjects. Together with the lack of a signifi-
cant main effect of cocoa and interaction effect, this result
suggests that theobromine is the major active compound in cocoa
that is responsible for the beneficial HDL-cholesterol–increasing
effect.
TABLE 4
Descriptive values of 24-h blood pressure and heart rate at baseline and end of intervention
1
Placebo Cocoa Theobromine Theobromine plus cocoa
n10 10 10 9
24-h SBP (mm Hg)
Baseline 120.7 618.6
2
117.0 615.4 120.8 615.7 119.2 616.2
4 wk 119.6 615.1 115.9 617.2 122.5 615.2 121.7 615.4
24-h DBP (mm Hg)
Baseline 75.1 613.3 75.3 612.2 75.1 610.9 75.2 611.4
4 wk 75.1 612.0 75.0 613.2 75.0 611.2 76.1 610.9
24-h heart rate (beats/min)
Baseline 68.9 68.0 71.0 612.8 70.7 610.1 68.8 611.1
4 wk 69.1 610.2 72.0 613.1 72.8 611.4 74.3 615.2
1
Data from all subjects for whom baseline and follow-up measurements were available were included. There were no
significant main or interaction effects for any variable (P.0.05; full factorial analysis with cocoa and theobromine as full
factorial parameters and sex and baseline values as covariates). DBP, diastolic blood pressure; SBP, systolic blood pressure.
2
Mean 6SD (all such values).
THEOBROMINE RAISES HDL CHOLESTEROL 1207
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
We thank all participants who took part in this study. We also thank the
clinical investigators Yves Donazzolo, Mathilde Latreille, and Ste
´phanie
Delcroix and the project manager Isabelle Blank at Eurofins Optimed for
the execution of the study throughout and for taking care of study partici-
pants. In addition, we thank our following colleagues at Unilever Research &
Development: Richard Draijer, Yuguang Lin, Mario Vermeer, and Winfried
Theis for their scientific input and critical review of the manuscript; Jamy
von Harras and Koos Scholtes for the development and production of test
products; Jeroen Sterken for coordinating the logistics of test products; and
Christian Gru
¨n, Monique van der Burg, and Ruud Poort for performing the
chemical analyses.
The authors’ responsibilities were as follows—NN, YEMPZ, EAHS, and
EAT: conception and design of study, analysis and interpretation of data, and
critical revision and final approval of the manuscript; EAHS, YEMPZ,
and EAT: blind review of data; and NN: drafting of the manuscript. All
authors are employed by Unilever Research & Development. All authors
are or were employed by Unilever R&D Vlaardingen at the time the research
was conducted. Unilever has no products enriched with theobromine under
development or on the market; however, it markets food products enriched
with plant sterols to lower LDL cholesterol. None of the authors had a con-
flict of interest.
REFERENCES
1. Buitrago-Lopea A, Sanderson J, Johnson L, Warnakula S, Wood A,
Franco OH. Chocolate consumption and cariometabolic disorders:
systematic review and meta-analysis. BMJ 2011;343:d4488.
2. Ried K, Sullivan T, Fakler P, Frank OR, Stocks NP. Does chocolate
reduce blood pressure? A meta-analysis. BMC Med 2010;8:8.
3. Ding EL, Hutfless SM, Ding X, Girotra S. Chocolate and prevention of
cardiovascular disease: a systematic review. Nutr Metab (Lond) 2006;
3:2.
4. Ferna
´ndez-Murga L, Tarin JJ, Garcia-Perez MA, Cano A. The impact
of chocolate on cardiovascular health. Maturitas 2011;69:312–21.
5. Jia L, Liu XA, Bai YY, Li SH, Sun K, He C, Hui RT. Short-term effect
of cocoa product consumption on lipid profile: a meta-analysis of
randomized controlled trials. Am J Clin Nutr 2010;92:218–25.
6. Shrime MG, Bauer SR, McDonald AC, Chowdhury NH, Coltart CEM,
Ding EL. Flavonoid-rich cocoa consumption affects multiple cardio-
vascular risk factors in a meta-analysis of short-term studies. J Nutr
2011;141:1982–8.
7. Tokede OA, Gaziano JM, Djousse L. Effects of cocoa products/dark
chocolate on serum lipids: a meta-analysis. Eur J Clin Nutr 2011;65:
879–86.
8. Baba S, Osakabe N, Kato Y, Natsume M, Yasuda A, Kido T, Fukuda K,
Muto Y, Kondo K. Continuous intake of polyphenolic compounds
containing cocoa powder reduces LDL oxidative susceptibility and has
beneficial effects on plasma HDL-cholesterol concentrations in hu-
mans’. Am J Clin Nutr 2007;85:709–17.
9. Khan N, Monagas M, Andres-Lacueva C, Casas R, Urpy
´
-Sarda M,
Lamuela-Raventos RM, Estruch R. Regular consumption of cocoa
powder with milk increases HDL cholesterol and reduces oxidized
LDL levels in subjects at high-risk of cardiovascular disease. Nutr
Metab Cardiovasc Dis 2012;22:1046–53.
10. Kris-Etherton PM, Derr JA, Mustad VA, Seligson FH, Pearson TA. Ef-
fects of a milk chocolate bar per day substituted for a high-carbohydrate
snack in young men on an NCEP/AHA Step 1 Diet. Am J Clin Nutr
1994;60(6 suppl):1037S–42S.
11. Mellor DD, Sathyapalan T, Kilpatrick ES, Beckett S, Atkin SL. High-
cocoa polyphenol-rich chocolate improves HDL cholesterol in type 2
diabetes patients. Diabet Med 2010;27:1318–21.
12. Mursu J, Voutilainen S, Nurmi T, Rissanen TH, Virtanen JK, Kaik-
konen J, Nyyssoonen K, Salonen JT. Dark chocolate consumption in-
creases HDL cholesterol concentration and chocolate fatty acids may
inhibit lipid peroxidation in healthy humans. Free Radic Biol Med
2004;37:1351–9.
13. Wang-Polagruto JF, Villablanca AC, Polagruto JA, Lee L, Holt RR,
Schrader HR, Ensunsa JL, Steinberg FM, Schmitz HH, Keen CL.
Chronic consumption of flavanol-rich cocoa improves endothelial
function and decreases vascular cell adhesion molecule in hype cho-
lesterolemic postmenopausal women. J Cardiovasc Pharmacol 2006;
47:S177–86.
14. Balzer J, Rassaf T, Heiss C, Kleinbongard P, Lauer T, Merx M,
Heussen N, Gross HB, Keen CL, Schroeter H, et al. Sustained benefits
in vascular function through flavanol-containing cocoa in medicated
diabetic patients - a double-masked, randomized, controlled trial. J Am
Coll Cardiol 2008;51:2141–9.
15. Crews WD, Harrison DW, Wright JW. A double-blind, placebo-controlled,
randomized trial of the effects of dark chocolate and cocoa on variables
associated with neuropsychological functioning and cardiovascular
health: clinical findings from a sample of healthy, cognitively intact
older adults. Am J Clin Nutr 2008;87:872–80.
16. Davison K, Coate AM, Buckley JD, Howe PRC. Effect of cocoa fla-
vanols and exercise on cardiometabolic risk factors in overweight and
obese subjects. Int J Obes (Lond) 2008;32:1289–96.
17. Engler MB, Engler MM, Chen CY, Malloy MJ, Browne A, Chiu EY,
Kwak HK, Milbury P, Paul SM, Blumberg J, et al. Flavonoid-rich dark
chocolate improves endothelial function and increases plasma epi-
catechin concentrations in healthy adults. J Am Coll Nutr 2004;23:
197–204.
18. Muniyappa R, Hall G, Kolodziej TL, Karne RJ, Crandon SK, Quon MJ.
Cocoa consumption for 2 wk enhances insulin-mediated vasodilatation
without improving blood pressure or insulin resistance in essential
hypertension. Am J Clin Nutr 2008;88:1685–96.
19. Taubert D, Roesen R, Lehmann C, Jung N, Schomig E. Effects of low
habitual cocoa intake on blood pressure and bioactive nitric oxide -
a randomized controlled trial. JAMA 2007;298:49–60.
20. Wan Y, Vinson JA, Etherton TD, Proch J, Lazarus SA, Kris-Etherton P.
Effects of cocoa powder and dark chocolate on LDL oxidative sus-
ceptibility and prostaglandin concentrations in humans. Am J Clin Nutr
2001;74:596–602.
21. NCEP Expert panel. Third Report of the National Cholesterol Educa-
tion Program (NCEP) Expert Panel on Detection, Evaluation, and
Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel
III) final report. Circulation 2002;106:3143–421.
22. Galleano M, Oteiza PI, Fraga CG. Cocoa, chocolate, and cardiovas-
cular disease. J Cardiovasc Pharmacol 2009;54:483–90.
23. Huxley RR, Neil HAW. The relation between dietary flavonol intake
and coronary heart disease mortality: a meta-analysis of prospective
cohort studies. Eur J Clin Nutr 2003;57:904–8.
24. Riksen NP, Smits P, Rongen GA. The cardiovascular effecs of meth-
ylxanthines. In: Fredholm BB, ed. Methylxanthines. Heidelberg, Ger-
many: Springer Verlag Berlin, 2011:413–37.
25. Smit HJ. Theobromine and the pharmacology of cocoa. In: Fredholm
BB, ed. Methylxanthines. Heidelberg, Germany: Springer Verlag
Berlin, 2011:201–34.
26. Graham I, Atar D, Borch-Johnsen K, Boysen G, Burell G, Cifkova R,
Dallongeville J, De Backer G, Ebrahim S, Gjelsvik G, et al. European
guidelines on cardiovascular disease prevention in clinical practice:
full text. Fourth Joint Task Force of the European Society of Cardi-
ology and other societies on cardiovascular disease prevention in
clinical practice (constituted by representatives of nine societies and by
invited experts). Eur J Cardiovasc Prev Rehabil 2007;14(suppl 2):S1–
113.
27. Singleton VL, Orthofer R, Lamuela-Raventos RM. Analysis of total
phenols and other oxidation substrates and antioxidants by means of
Folin-Ciocalteu reagent. Oxidants and Antioxidants Pt A 1999;299:
152–78.
28. Ptolemy AS, Tzioumis E, Thomke A, Rifai S, Kellogg M. Quantifi-
cation of theobromine and caffeine in saliva, plasma and urine via
liquid chromatography-tandem mass spectrometry: a single analytical
protocol applicable to cocoa intervention studies. J Chromatogr B
Analyt Technol Biomed Life Sci 2010;878:409–16.
29. Van Gaal LF, Mertens IL, Ballaux D. What is the relationship between
risk factor reduction and degree of weight loss? Eur Heart J Suppl
2005;7:L21–6.
30. Fraga CG, Actis-Goretta L, Ottaviani JI, Carrasquedo F, Lotito SB,
Lazarus S, Schmitz HH, Keen CL. Regular consumption of a flavanol-
rich chocolate can improve oxidant stress in young soccer players. Clin
Dev Immunol 2005;12:11–7.
31. Grassi D, Lippi C, Necozione S, Desideri G, Ferri C. Short-term ad-
ministration of dark chocolate is followed by a significant increase in
insulin sensitivity and a decrease in blood pressure in healthy persons.
Am J Clin Nutr 2005;81:611–4.
32. Grassi D, Necozione S, Lippi C, Croce G, Valeri L, Pasqualetti P,
Desideri G, Blumberg JB, Ferri C. Cocoa reduces blood pressure and
1208 NEUFINGERL ET AL
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
insulin resistance and improves endothelium-dependent vasodilation in
hypertensives. Hypertension 2005;46:398–405.
33. Osakabe N, Baba S, Yasuda A, Iwamoto T, Kamiyama M, Tokunaga T,
Kondo K. Dose-response study of daily cocoa intake on the oxadative
susceptibility of low-density lipoprotein in healthy human volunteers. J
Health Sci 2004;50:679–84.
34. Bruckert E. Epidemiology of low HDL-cholesterol: results of studies
and surveys. Eur Heart J Suppl 2006;8:F17–22.
35. Kodama S, Tanaka S, Saito K, Shu M, Sone Y, Onitake F, Suzuki E,
Shimano H, Yamamoto S, Kondo K, et al. Effect of aerobic exercise
training on serum levels of high-density lipoprotein cholesterol -
a meta-analysis. Arch Intern Med 2007;167:999–1008.
36. Durstine JL, Grandjean PW, Davis PG, Ferguson MA, Alderson NL,
Dubose KD. Blood lipid and lipoprotein adaptations to exercise -
a quantitative analysis. Sports Med 2001;31:1033–62.
37. Katcher HI, Hill AM, Lanford JLG, Yoo JS, Kris-Etherton PM. Life-
style approaches and dietary strategies to lower LDL-cholesterol and
triglycerides and raise HDL-cholesterol. Endocrinol Metab Clin North
Am 2009;38:45.
38. Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP,
Knoke JD, Jacobus DR, Bangdivala S, Tyroler HA. High-density li-
poprotein cholesterol and cardiovascular disease. Four prospective
American studies. Circulation 1989;79:8–15.
39. Birjmohun RS, Hutten BA, Kastelein JJ, Stroes ES. Efficacy and safety
of high-density lipoprotein cholesterol-increasing compounds: a meta-
analysis of randomized controlled trials. J Am Coll Cardiol 2005;45:
185–97.
40. Bhatt KN, Wells BJ, Sperling LS, Baer JT. High-density lipoprotein
therapy: is there hope? Curr Treat Options Cardiovasc Med 2010;12:
315–28.
41. Fruchart JC, Sacks F, Hermans MP, Assmann G, Brown WV, Ceska R,
Chapman MJ, Dodson PM, Fioretto P, Ginsberg HN, et al. The residual
risk reduction initiative: a call to action to reduce residual vascular risk
in patients with dyslipidemia. Am J Cardiol 2008;102:1K–34K.
42. Abbott RD, Wilson PWF, Kannel WB, Castelli WP. High-density
lipoprotein cholesterol, total cholesterol screening, and myocardial-
infarction - the Framingham-Study. Arteriosclerosis 1988;8:207–11.
43. Wilson PWF, D’Agostino RB, Levy D, Belanger AM, Silbershatz H,
Kannel WB. Prediction of coronary heart disease using risk factor
categories. Circulation 1998;97:1837–47.
44. NEVO Foundation. Dutch food composition table 2001. The Hague,
Netherlands: PlantijnCasparie, 2011.
THEOBROMINE RAISES HDL CHOLESTEROL 1209
at UNILEVER on October 26, 2015ajcn.nutrition.orgDownloaded from
