Current Nutrition Reports

, Volume 2, Issue 4, pp 267–273

Effects of Dark Chocolate and Cocoa Products on Endothelial Function: A Meta-Analysis

  • Andrew B. Petrone
  • J. Michael Gaziano
  • Luc Djoussé
Cardiovascular Disease (L Djoussé, Section Editor)

DOI: 10.1007/s13668-013-0058-y

Cite this article as:
Petrone, A.B., Gaziano, J.M. & Djoussé, L. Curr Nutr Rep (2013) 2: 267. doi:10.1007/s13668-013-0058-y

Abstract

Consumption of dark chocolate, a rich source of flavonoids, has been associated with reduced risk of cardiovascular disease. However, underlying pathophysiologic mechanisms are not fully elucidated. We reviewed existing evidence on the effect of cocoa consumption on flow-mediated dilation (FMD) by conducting a literature search using PubMed and Embase for completed, randomized, controlled trials. The primary effect measure was the difference in means of the final measurement between the intervention and control groups. Nineteen clinical trials with a total of 454 participants were included. Treatment duration ranged from 2 hours to 12 weeks. Pooled estimate showed that intervention with dark chocolate significantly increased FMD levels by 2 % (95 % confidence interval 1.6-2.39 %) compared with placebo/control group. Similar results were seen when stratified by study design, geographic location, cocoa dose, or study quality. In addition, the effect size was greater in individuals with cardiovascular risk factors. In summary, intervention with cocoa improved endothelial function as measured by FMD.

Keywords

Dark chocolate Cocoa Flavonoids Endothelial function Flow-mediated dilation 

Introduction

Cardiovascular disease (CVD) affects approximately 83.6 million adults in the United States and accounts for 1 in 3 deaths, making CVD the leading cause of death in the United States [1••]. In 2009, estimated costs of CVD were $300 billion, with total direct costs projected to increase to $1.5 trillion by 2030 [1••]. Thus, identifying cost-effective intervention strategies to reduce CVD morbidity and mortality is an important public health goal.

Previous studies have suggested that high consumption of dark chocolate and cocoa is associated with reduced risk of CVD and cardiovascular mortality [2•, 3••, 4]. Additionally, prospective cohort studies have shown an inverse relationship between cocoa consumption and cardiovascular mortality [5, 6]. It has been suggested that some of the benefits of cocoa products on CVD may be due to favorable effects of cocoa products on lipid profiles [7••] and blood pressure [8].

Cocoa products are rich is plant phytochemicals, especially flavonoids, with strong antioxidant properties [9]. Because endothelial dysfunction plays a critical role in the development and progression of atherosclerosis [10], it is possible that consumption of chocolate or other cocoa products may improve endothelial function. Experiments with endothelial cell cultures have shown that flavonoids reduce the down-regulation of nitric oxide synthase, thereby promoting the production of nitric oxide and help maintain normal endothelial function [11, 12].

Currently, it is not very clear whether cocoa products, including dark chocolate consumption, consistently promote normal endothelial function. Therefore, in this meta-analysis of randomized, controlled trials, we reviewed the effects of dark chocolate/cocoa products on FMD.

Materials and Methods

Search Strategy and Study Selection

We searched PubMed and Embase to identify randomized, controlled trials that examined the effects of dark chocolate, cocoa containing beverages, or high flavanol diets on FMD. For PubMed search, we used MeSH (medical subject heading) terms: chocolate, flavonol, flavonoid, cocoa, cacao, endothelial, and endothelium. We also used tiab (title or abstract) mentions of chocolate, flavonol, flavonoid, cocoa, cacao, endothelial, endothelium, flow mediated dilation, brachial artery diameter, brachial artery ultrasound, and FMD. Terms from PubMed search were used in Embase search without any restrictions. We excluded articles that were: a) not published in English language; b) not relevant to the topic; c) not randomized controlled trials; d) animal studies; e) review articles; and f) conference abstracts. We also manually searched reference lists of all relevant articles to identify pertinent studies. Retrieved studies were included if they met the following criteria: a) investigation of flavonoid-rich cocoa products; b) randomized, controlled parallel-arm or crossover design; c) studied subjects aged 18+ years; and d) available “baseline” and “postintervention” mean and standard deviation for FMD(%). Additionally, we excluded studies in which flavonoid-rich cocoa was mixed with other interventions.

In case of incomplete data in selected manuscripts or when only figures were published, corresponding authors were contacted in order to obtain actual means and standard deviations of FMD at baseline and postintervention.

Data Extraction and Quality Assessment

Data were extracted by the lead author. Extracted data included study characteristics (first author’s name; year of publication; sample size, mean age, and characteristics of participants; study design; polyphenol amounts and dietary intervention in the active and control arm; duration of the study; and geographic location of the trial) and information on baseline and final FMD. Study quality was evaluated using Jadad score [13]. The maximum possible score was 5.

Data Synthesis and Statistical Analysis

The effect size used in our analyses was the difference in mean FMD at the end of each intervention between intervention and placebo. In one study, standard error was converted to standard deviation [14]. Sensitivity analysis was conducted by stratifying studies according to duration, design, geographic location, health status, cocoa dose, and study quality.

Heterogeneity across studies was assessed by Cochrane Q test; p < 0.1 was considered statistically significant for heterogeneity. The magnitude of heterogeneity was evaluated by I2 statistic (percentage of the variability in effect estimate that is due to heterogeneity rather than sampling error). In circumstances where the test of heterogeneity was statistically significant, the pooled estimate was calculated using random effect model. Data synthesis and statistical analyses were completed using Cochrane Collaboration Review Manager (RevMan software version 5.2; Cochrane Collaboration, Oxford, UK).

Results

Of the 315 records screened, we analyzed data from 19 randomized, controlled trials with a total of 454 individuals [14, 15, 16, 17, 18, 19, 20, 21••, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32•]. Four of the trials provided multiple effect measurements because they either included multiple populations [20, 27] or presented both acute and long-term effects of cocoa products [14, 19]. Characteristics of the trials included in the current report are shown in Table 1. Study duration ranged from 2 hours to 12 weeks. Ten of the trials were conducted on healthy individuals, whereas ten used subjects with at least one CVD risk factor (1 trial on diabetic individuals [15], two trials on overweight or obese individuals [14, 16], one trial on individuals diagnosed with coronary artery disease [24], one trial on heart failure patients [19], one trial of hypertensive subjects [20], one trial in hypercholesterolemic subjects [31], and three trials of smokers [23, 25, 27]). Eleven of the trials had a Jadad score ≥3 [14, 15, 16, 17, 19, 23, 24, 28, 29, 31, 32•].
Table 1

Characteristics of the 19 trials meta-analyzed

Study

Intervention

Control

Flavonol intake (active arm)

Flavonol intake (control)

Design

Jadad score

Duration

Population characteristics

# of subjects

Age*

Balzer et al. (2008) [15]

High flavonol cocoa drink

Low flavonol cocoa drink

963 mg flavonol

75 mg flavonol

Randomized parallel group

5

30 days

Medicated diabetic

41

Intervention: 63.1 ± 8.3

Control:

64.4 ± 8.6

Berry et al. (2010) [16]

High flavonol cocoa drink

Low flavonol cocoa drink

701 mg flavonol

22 mg flavonol

Randomized crossover

4

2 hours

Overweight and obese

21

54.9 ± 2.2

Davison et al. (2008) [14]

High flavonol cocoa drink

Low flavonol cocoa drink

902 mg flavonol

36 mg flavonol

Randomized parallel group

4

2 hours, 12 weeks

Overweight and obese

49

44.4 ± 4.4 -45.5 ± 4.0 across treatment groups

Engler et al. (2004) [17]

46 g dark chocolate

46 g low-flavanoid dark chocolate

213 mg procyanidins

Trace amounts

Randomized parallel group

4

2 weeks

Healthy

21

Intervention:

31.8 ± 3.2

Control:

32.5 ± 2.9

Faridi et al. (2008) [18]

74 g dark chocolate

Placebo

821 mg procyanidins

0 mg procyanidins

Randomized crossover

2

2 hours

Healthy

44

52.8 ± 11

Flammer et al. (2012) [19]

40/80 g dark chocolate

Cocoa-free chocolate

15.6/31.2 epicatechin equivalents/g

0 epicatechin equivalents/g

Randomized parallel group

4

2 hours, 4 weeks

Congestive heart failure

20

Intervention:

60.3 ± 10.1

Control:

58.1 ± 11.9

Grassi et al. (2005) [20]

100 g dark chocolate

90 g white chocolate

88 mg flavonols

0 mg flavonols

Randomized crossover

1

15 days

Never treated essential hypertension and healthy

20 hypertensive, 15 healthy

Hypertensive: 43.7 ± 7.8

Healthy:

33.9 ±7.6

Grassi et al. (2012) [21••]

100 g dark chocolate

100 g white chocolate

447 mg epicatechin

Trace amounts

Randomized crossover

1

3 days

Healthy

12

28.2 ± 2.7

Heiss et al. (2003) [22]

High flavanol cocoa drink

Low flavonol cocoa drink

176 mg flavonols

<10 mg flavonols

Randomized crossover

2

2 hours

At least one cardiovascular risk factor

20

41 ± 14

Heiss et al. (2005) [23]

High flavonol cocoa drink

Low flavonol cocoa drink

176-185 mg flavonol

<11.5 mg flavonol

Randomized crossover

4

2 hours

Healthy smokers

11

31 ± 1

Heiss et al. (2010) [24]

High flavonol cocoa drink

Low flavonol cocoa drink

750 mg flavonol

18 mg flavonol

Randomized crossover

3

4 weeks

Coronary artery disease

16

64 ± 3

Hermann et al. (2006) [25]

40 g dark chocolate

40 g white chocolate

n/a

n/a

Randomized parallel group

1

2 hours

Male smokers

20

n/a

Kim et al. (2012) [26]

32 g dark chocolate

No chocolate

n/a

n/a

Randomized crossover

1

3 days

Healthy females

20

27.1 ± 3.5

Loffredo et al. (2011) [27]

40 g dark chocolate

40 g milk chocolate

n/a

n/a

Randomized crossover

2

2 hours

Smokers and healthy controls

20 smokers, 20 healthy

Healthy:

33 ± 11

Smokers:

33 ± 11

Monahan et al. (2011) [28]

Cocoa drink

Placebo

146.0 mg flavonols

0 mg flavonols

Randomized crossover

3

2 hours

Healthy

23

63 ± 2

Njike et al. (2011) [29]

Cocoa drink without sugar

Placebo

805 mg flavonols

9 mg flavonols

Randomized crossover

3

6 weeks

Healthy

37

51.9 ± 10.8

Vlachopoulos et al. (2005) [30]

100 g dark chocolate

Sham-eating

2620 mg procyanidins

0 mg flavonols

Randomized crossover

1

2 hours

Healthy

17

28.9 (range = 24-32)

Wang-Polagruto et al. (2006) [31]

High flavonol cocoa drink

Low flavonol cocoa drink

446 mg flavonol

43 mg flavonol

Randomized parallel group

4

6 weeks

Postmenopausal hypercholesterolemic females

17

Intervention:

57.7 ± 2.2

Control:

55.4 ± 1.7

Westphal et al. (2011) [32•]

Fatty meal with cocoa

Flavonol poor fatty meal

918 mg flavonol

14.63 mg flavonol

Randomized crossover

4

2 hours

Healthy

18

25.2 ± 2.5

*Mean ± standard deviation

Mean age of study participants was 46.9 years. The difference in FMD after intervention with dark chocolate or cocoa compared to a placebo was 2 % (95 % confidence interval [CI]: 1.6-2.39 %) in the pooled analysis using random-effect model (I2: 82 %; Fig. 1). Subgroup analysis stratified by duration of the intervention, study design, participants’ health status, and study quality is shown in Table 2. Trials that lasted less than one day showed a significant increase in FMD (2.25 %; 95 % CI: 1.69-2.81 %), whereas trials that were long-term (3 days to 12 weeks) had a comparable effect size (1.76 %; 95 % CI: 1.12-2.4 %). Stratification by health status showed a slightly greater effect of cocoa products in individuals with one or more CVD risk factors compared with healthy individuals (2.36 % vs. 1.53 %). Study design, geographic location, cocoa dose, and exclusion of lower quality studies did not alter results (Table 2).
Fig. 1

Meta-analysis of the effect of chocolate/cocoa consumption on flow-mediated dilation. IV, inverse variance; random, random-effects model. Values are % flow-mediated dilation. Davison a: 2 hours; Davison b: 12 weeks; Flammer a: 2 hours; Flammer b: 4 weeks; Grassi a: hypertensive subjects; Grassi b: healthy subjects; Loffredo a: healthy subjects; Loffredo b: smokers

Table 2

Effects of cocoa products/chocolate on flow-mediated dilation stratified by duration, study design, health status, and other factors

Variables

Number of effect sizes

FMD (%) mean difference (95 % CI)

I2 (I-squared)

Duration

 Short term (<1 day)

12

2.25 (1.69-2.81)

86 %

 Long term (>1 day)

11

1.76 (1.12-2.4)

78 %

Design

 Crossover

15

2.02 (1.54-2.5)

85 %

 Parallel arm

8

2.05 (1.18-2.93)

71 %

Health status

 Healthy

10

1.53 (1.05-2.02)

81 %

 One or more CVD risk factors

13z

2.36 (1.82-2.89)

68 %

Chocolate dose

 <50 g chocolate

6

1.97 (0.98-2.95)

55 %

Study location

 Europe

14

1.87 (1.3-2.45)

88 %

 Other

9

1.94 (1.75-2.14)

36 %

Study quality

 Jadad score ≥3

13

2.09 (1.6-2.58)

82 %

Number of effect sizes does not add to 19 because some studies contributed multiple effect sizes: Davison et al., Flammer et al. measured both acute and long term follow-up; Grassi 2005 et al., Loffredo et al. included both healthy individuals and subjects with a CVD risk factor

Discussion

In this meta-analysis of randomized, controlled trials, we found a statistically significant increase in FMD following an intervention with dark chocolate or cocoa products. In secondary analyses, the effect of dark chocolate/cocoa on FMD was strongest in trials that studied participants with one or more CVD risk factors.

These results are consistent with previous meta-analyses, which found that chocolate consumption significantly improved FMD by 3.99 % after acute intake, 1.53 % across all studies, and 3.19 % and 1.34 % for acute and long-term consumption, respectively [33, 34, 35]. Our finding of a slighter greater effect of chocolate/cocoa on FMD in individuals with one or more CVD risk factors is important, because previous research suggested that a 1 % increase in FMD is associated with a 13 % reduction in risk of cardiovascular events [36].

The observed increase in FMD after cocoa/dark chocolate consumption may be due to flavonoid content in the experimental groups, which ranged from 88 to 963 mg. Flavonoids in cocoa, such as epicatechin, appears to improve nitric oxide (NO) synthesis in the endothelium with subsequent vasodilation [37]. NO modulates endothelial cell activity and is produced by endothelial nitric oxide synthase (eNOS). Treatment of cell with (--)-epicatechin activates eNOS and stimulates NO production with subsequent increase in vasodilation [11]. In a study of circulating flavonoids, epicatechin predicted the magnitude of FMD (R2 for the multivariate model was 0.31), suggesting that epicatechin is directly involved in the improvements in vascular function observed with cocoa intervention [38]. These beneficial effects of epicatechin are reduced by administration of eNOS inhibitors [38, 39].

This study has several strengths. First, a large sample size affords more precise estimates of the effect measure than those derived from individual small studies. Second, we only included randomized, controlled trials in order to minimize confounding by unknown and unmeasured factors. Finally, compared with previous meta-analyses conducted on this topic, we had adequate sample size to further conduct stratified analyses by key factors.

Our study also has several limitations. A majority of the trials used quantities of cocoa that are larger than the usual amount of cocoa or chocolate consumed in the general population. This makes the translation of current findings difficult as excess calories from chocolate may lead to weight gain with resulting adverse health effects. Second, because the longest intervention lasted 12 weeks, it is unclear whether sustained chocolate consumption over longer timeline (several months or years) would yield similar benefits. A wide variety in the formulation of cocoa products used in reviewed trials makes it difficult to recommend the most effective cocoa product (type, ingredients, and quantity). Finally, FMD is a nonstandardized measurement of endothelial function that can vary depending on reader, cuff placement, room temperature, and administration [40, 41]. However, error in FMD measurement would tend to bias our estimates towards the null. If such hypothesis were true, then the true effect of cocoa/chocolate product would be higher than the observed effect size. Unfortunately, we only identified two studies using peripheral artery tonometry (PAT), which is independent of the examiner to assess the effects of chocolate/cocoa on endothelial function [42, 43••].

Conclusions

Overall, the current meta-analysis is consistent with a beneficial effect of dark chocolate and cocoa products on endothelial function. Future studies are needed to determine optimal amount of chocolate/cocoa, frequency of usage, flavonoid content, type of flavonoid that yields the largest effect size, and other patterns of cocoa consumption that could provide health benefits while minimizing any adverse effects.

Acknowledgments

We thank Drs Isabella Sudano, Roberto Corti, Kate Davison, Peter Howe, Narelle Berry, Mary Engler, David Katz, Valentine Njike, Lorenzo Loffredo, Konstantinos Aznaouridis, Christian Heiss, and Michelle Pinto-Evans for providing detailed information on their FMD measurements upon our request.

Compliance with Ethics Guidelines

Conflict of Interest

Andrew B. Petrone declares that he has no conflict of interest.

J. Michael Gaziano has no conflicts to declare.

Luc Djoussé has received travel reimbursement to attend the International Society for Heart Research meeting (San Diego, CA) in July 2013 from The Hershey Company.

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.

Copyright information

© Springer Science+Business Media New York 2013

Authors and Affiliations

  • Andrew B. Petrone
    • 1
  • J. Michael Gaziano
    • 1
    • 2
    • 3
  • Luc Djoussé
    • 1
    • 2
    • 3
  1. 1.Division of Aging, Department of MedicineBrigham and Women’s HospitalBostonUSA
  2. 2.Massachusetts Veterans Epidemiology and Research Information Center (MAVERIC)Boston Veterans Affairs Healthcare SystemBostonUSA
  3. 3.Geriatric Research (GRECC)Boston Veterans Affairs Healthcare SystemBostonUSA

Personalised recommendations