Clinical Science

Research article

Acute and chronic effects of flavanol-rich cocoa on vascular function in subjects with coronary artery disease: a randomized double-blind placebo-controlled study

H. M. Omar Farouque, Michael Leung, Sarah A. Hope, Mauro Baldi, Clyde Schechter, James D. Cameron, Ian T. Meredith


Evidence suggests that flavonoid-containing diets reduce cardiovascular risk, but the mechanisms responsible are unclear. In the present study, we sought to determine the effect of flavanol-rich cocoa on vascular function in individuals with CAD (coronary artery disease). Forty subjects (61±8 years; 30 male) with CAD were recruited to a 6-week randomized double-blind placebo-controlled study. Subjects consumed either a flavanol-rich chocolate bar and cocoa beverage daily (total flavanols, 444 mg/day) or matching isocaloric placebos daily (total flavanols, 19.6 mg/day) for 6 weeks. Brachial artery FMD (flow-mediated dilation) and SAC (systemic arterial compliance) were assessed at baseline, 90 min following the first beverage and after 3 and 6 weeks of daily consumption. Soluble cellular adhesion molecules and FBF (forearm blood flow) responses to ACh (acetylcholine chloride; 3–30 μg/min) and SNP (sodium nitroprusside; 0.3–3 μg/min) infusions, forearm ischaemia and isotonic forearm exercise were assessed at baseline and after 6 weeks. FMD, SAC and FBF responses did not differ between groups at baseline. No acute or chronic changes in FMD or SAC were seen in either group. No difference in soluble cellular adhesion molecules, FBF responses to ischaemia, exercise, SNP or ACh was seen in the group receiving flavanol-rich cocoa between baseline and 6 weeks. These data suggest that over a 6-week period, flavanolrich cocoa does not modify vascular function in patients with established CAD.

  • blood flow
  • coronary disease
  • diet
  • endothelium
  • flavonoid


Cocoa and chocolate are plant foods derived from the cacao bean (Theobroma cacao), which are rich in flavonoids. The main subclasses of flavonoids found in cocoa and chocolate are the flavanol monomers catechin and epicatechin, and oligomers derived from these monomeric units known as procyanidins [1]. Chocolate and cocoa have among the highest flavonoid contents of all foods on a per weight basis and appear to be a significant contributor to the total dietary intake of flavonoids [2,3].

There is evidence from population-based studies to suggest that a diet high in flavonoids may be associated with reduced cardiovascular risk [47]. Although the mechanism of benefit is unclear, flavonoids have a number of properties that may contribute to cardiovascular protection, including antioxidant and antiplatelet activity, immunoregulatory properties and beneficial effects on the endothelium [1,812]. The latter property is of potential clinical relevance, since endothelial dysfunction is an early event in the process of atherogenesis and one that may contribute to ischaemic sequelae [13,14]. Dietary intervention to reverse endothelial dysfunction at either an early stage or once atherosclerosis was present could conceivably reduce the burden of complications that result from this disease [1517]. Hence the aim of the present study was to determine the effects of flavanol-rich cocoa on invasive and non-invasive measurements of endothelial and vascular function over a 6-week period in patients with chronic ischaemic heart disease using a randomized double-blind placebo-controlled study design.


Study population

Forty subjects with angiographically documented CAD (coronary artery disease; >50% stenosis in at least one epicardial coronary artery) were recruited for the study. All subjects underwent screening by history, physical examination, haematological and biochemical analyses and ECG. Subjects had been clinically stable for at least 3 months before study enrolment with no modifications made to medication. Exclusion criteria included age <18 or >80 years, uncontrolled diabetes mellitus or significant non-cardiac medical illnesses. The Southern Health Human Research Ethics Committee approved the study, and all subjects gave written informed consent.

Study protocol

A randomized double-blind placebo-controlled trial was performed. After initial screening, subjects received, in random order, both a flavanol-rich chocolate bar and cocoa beverage daily (flavanol group: 444 mg of flavanols daily; ≈107 mg of epicatechin monomer daily) or matching isocaloric placebo daily [placebo (non-flavanol) group: 19.6 mg of flavanols daily; ≈4.7 mg of epicatechin monomer daily] for 6 weeks. The chocolate bar (48 g/serving) contained approx. 220 kcal (1 kcal≈4.184 kJ), and its constituents included ≈33% fat, ≈4% protein, ≈56% carbohydrate, ≈0.1% caffeine and ≈0.6% theobromine. The cocoa beverage (18 g/serving) contained approx. 55 kcal and was composed of ≈10–12% fat, ≈19% protein, ≈50% carbohydrate, ≈0.2% caffeine and ≈2% theobromine. Each packet was mixed thoroughly with 100 ml of warm water without the addition of milk. The placebo chocolate bar and beverage had the same macronutrient, caffeine and theobromine content as the flavanol-rich cocoa products.

Randomization was achieved with the aid of a computer-generated random numbers program. Study products were provided to subjects in unmarked silver wrappers. All study measurements were performed at baseline and at 6 weeks (see below). To examine for acute effects, conduit vessel endothelial function was assessed before and 90 min after the consumption of the first test meal (either a single flavanol-rich cocoa beverage or isocaloric placebo). An interim assessment of conduit vessel endothelial function and SAC (systemic arterial compliance) was performed at 3 weeks. Subjects underwent vascular studies in the fasting state (with the exception of the cocoa beverage used in the acute 90 min study) and at the same time in the morning to minimize the effect of diurnal fluctuations in vascular reactivity [18]. Vasoactive drugs were withheld on the morning of the procedure. Clinical review was performed on a weekly basis to ensure compliance with treatment. Subjects were advised to maintain their usual diet throughout the study period.

FMD (flow-mediated dilation)

FMD is dependent on the release of NO (nitric oxide) [19] and has been widely used as a measure of conduit vessel endothelial vasodilator function [8,15,16]. Brachial artery FMD was measured using an ultrasound machine with a 7–10-MHz linear array transducer (HDI Ultramark 9; ATL), as described previously [20]. Recordings on to super VHS videotape were made at baseline, during reactive hyperaemia [induced by inflating a BP (blood pressure) cuff on the forearm to 200 mmHg for 5 min], after 15 min of rest and following 0.4 mg of sublingual GTN (glyceryl trinitrate) administration. Two representative frames from each time point (baseline, 45–60 s after cuff deflation and before and 3–5 min after GTN) were subsequently digitized and analysed. Each frame was measured three times, and the average of all measurements for each time point was calculated. Endothelium-dependent FMD was calculated as the percentage change in brachial artery diameter in reference to the resting state. Endothelium-independent vasodilation was assessed after GTN administration. Brachial artery reactivity was assessed at baseline, 90 min after the first test meal and at 3 and 6 weeks. In our laboratory, the intra-observer and inter-observer variability in measuring brachial artery diameter (mean±S.D. of the absolute difference) was 0.04±0.03 mm (average coefficient of variation=0.76%) and 0.12±0.11 mm (average coefficient of variation=1.9%) respectively.

VOP (venous occlusion plethysmography)

Strain-gauge VOP (D. E. Hokanson) was performed in a quiet temperature-controlled vascular research laboratory, as described previously [21,22]. ACh (acetylcholine chloride; Miochol; Iolab Pharmaceuticals) and SNP (sodium nitroprusside; David Bull Laboratories) were infused into the brachial artery in turn, and FBF (forearm blood flow) was measured to assess endothelium-dependent and -independent vasodilation respectively. Graded doses of ACh at 3, 10 and 30 μg were infused for 5 min, each at 0.4 ml/min. After allowing blood flow to return to baseline, graded doses of SNP were infused at 0.3, 1 and 3 μg for 5 min each. FBF was measured over a 2-min period after each drug dose had been infused for 3 min. FBF was also assessed in response to 2 min of wrist flexion–extension exercise (functional hyperaemic blood flow) and 5 min of forearm ischaemia induced by upper arm cuff occlusion of the brachial artery (reactive hyperaemia). The time course of vasodilation immediately after exercise and ischaemia was assessed by describing the peak and sustained hyperaemic responses (volumes repaid at 1 and 5 min) [21,22]. VOP was performed at baseline and at 6 weeks.


SAC was measured non-invasively with an ultrasound technique [23,24]. In brief, flow velocity in the ascending aorta was measured from the suprasternal notch using a continuous-wave hand-held Doppler flow velocimeter (Multi Dopplex II; Huntleigh Technology). Aortic root dimensions were assessed by two-dimensional echocardiography (HDI Ultramark 9; ATL). Volumetric flow was then calculated from the product of average systolic aortic flow velocity and aortic root area. Aortic root pressure was estimated by applanation tonometry of the right carotid artery (Millar Mikro-Tip pressure transducer model SSD-713; Millar Instruments). Systemic arterial compliance was calculated from the formula Ad/[R(PsPd)], where Ad is diastolic area, R is total peripheral resistance and Ps and Pd are end-systolic and end-diastolic BPs respectively [23,24]. Each subject was studied in the supine position at baseline, 90 min after the first test meal and at 3 weeks and 6 weeks after study commencement.

Biochemical analyses

Circulating biomarkers of endothelial function, including soluble ICAM-1 (intercellular cell-adhesion molecule-1), VCAM-1 (vascular cell adhesion molecule-1), P-selectin and E-selectin, were assayed in duplicate using commercially available ELISA kits (R&D Systems). High-sensitivity CRP (C-reactive protein) was measured using a particle-enhanced turbidimetric immunoassay technique (Dade Behring). Assessment of oxidized LDL (low-density lipoprotein) was by the Mercodia Oxidized LDL ELISA assay. Other blood tests included total cholesterol, serum triacylglycerols, LDL, HDL (high-density lipoprotein), lipoprotein (a), serum glucose and plasma insulin levels. These analyses were performed at baseline and at 6 weeks in the Department of Clinical Biochemistry at Monash Medical Centre. Plasma epicatechin concentrations were measured using HPLC by an external laboratory. Blood samples for epicatechin determinations were taken at baseline and at 90 min after the first cocoa beverage.

Statistical analysis

Baseline subject characteristics are expressed as means±S.D., and other data as means±S.E.M. Continuous variables between groups were compared using the Mann–Whitney U test, and categorical variables were compared with the Pearson χ2 test. Generalized estimating equations were used to analyse the repeated FMD, VOP and SAC measurements at baseline, after the first test meal and at 3 and 6 weeks. A P value <0.05, without adjustment for multiple comparisons, was considered statistically significant. All analyses were performed using Stata version 8.2 (StataCorp LP).

Forty patients were chosen for the present study based on the expectation of a 4–5% improvement in FMD with dietary intervention [16]. Our sample size calculation indicated that between 16 and 20 subjects would be required in each treatment arm to detect a difference at a P value <0.01 with 90% power.


Subject characteristics are shown in Table 1. There were no differences in baseline demographic and morphometric characteristics. The group receiving flavanol-rich cocoa had higher baseline total cholesterol (P=0.013; see Table 4) and LDL-cholesterol (P=0.005; see Table 4) compared with the placebo (non-flavanol) group. Nineteen subjects (95%) enrolled in each arm completed the study. The duration of treatment in each group was similar (42.3±1.7 days in the flavanol group compared with 42.7±1.0 days in the placebo group; P=0.42). One subject (5%) in each arm dropped out after the acute FMD study. There was no change in weight during the study (0.40±0.35 kg in the flavanol group compared with 0.68±0.49 kg in the placebo group; P=0.487). Plasma epicatechin concentration increased acutely in all subjects receiving the flavanol-rich cocoa beverage, but in only five subjects receiving the placebo beverage (mean increase of 153.7 nmol/l in the flavanol group compared with 2.9 nmol/l in the placebo group; P<0.0001). Treatment was well tolerated by the subjects who completed the study, and compliance was excellent as judged by wrapper count and patient reports. No patient suffered an adverse cardiovascular event.

View this table:
Table 1 Baseline characteristics of the two groups

Values are means±S.E.M., or n (%). CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention.

Conduit vessel endothelial function

There were no differences in the change in brachial artery diameter and blood flow induced by post-ischaemic hyperaemia and GTN at baseline (Figure 1 and Table 2).

Figure 1 Changes in brachial artery diameter and blood flow to FMD and GTN at baseline (pre-cocoa) in the flavanol and non-flavanol (placebo) groups
View this table:
Table 2 Relative changes in brachial artery diameter and blood flow to FMD and GTN in the flavanol group compared with the placebo (non-flavanol) group at baseline, after 90 min (acute) and after 3 and 6 weeks

Values are means (95% CI).

There was no significant change in FMD or GTN-induced vasodilation at 90 min post-beverage consumption (Figure 2 and Table 2).

Figure 2 Changes in brachial artery diameter and blood flow to FMD and GTN at 90 min after acute ingestion of a non-flavanol (placebo)- or flavanol-containing cocoa beverage

No further differences were observed at 3 or 6 weeks following daily chocolate and cocoa beverage consumption (Figure 3 and Table 2).

Figure 3 Changes in brachial artery diameter and blood flow to FMD and GTN in patients ingesting non-flavanol (placebo)- or flavanol-containing cocoa at baseline, after 90 min (acute) and after 3 and 6 weeks

Forearm microvascular endothelial function

Similar to the FMD response, the evolution over time of the endothelium-independent (SNP) FBF dose–response curves showed no significant difference averaged over the doses of SNP [change in FBF at 6 weeks in the flavanol group compared with the placebo group=−0.53 (95% CI, −1.28 to +0.22), where CI is confidence interval]. However, the endothelium-dependent (ACh) FBF dose–response curves differed slightly between the flavanol-rich cocoa and placebo groups averaged over doses of ACh (Figure 4) [change in FBF at 6 weeks in the flavanol group compared with the placebo group=−1.61 (95% CI, −2.78 to −0.4268)].

Figure 4 Dose–response curves to ACh and SNP for FBF at baseline (pre-cocoa) and after 6 weeks of ingesting non-flavanol (placebo)- or flavanol-containing cocoa (post-cocoa)

*P<0.05. mcg, μg.

Relatedly, there was a significant difference at the highest concentrations of ACh between baseline and 6 weeks in the placebo group [change in FBF at 6 weeks=1.44 (95% CI, +0.13 to +2.75) and 1.92 (95% CI, +0.62 to +3.23) at doses of 10 and 30 μg/min ACh respectively]. At 6 weeks, forearm hyperaemic responses following a brief period of exercise or ischaemia were not altered by the ingestion of flavanol-rich cocoa compared with the baseline state (Table 3).

View this table:
Table 3 Effect of cocoa on FBF after 6 weeks

Values are means±S.E.M. FHBF, functional hyperaemic blood flow; RHBF, reactive hyperaemic blood flow. P values are not significant.

SAC and haemodynamics

There was no difference in systolic, diastolic or mean arterial pressure, heart rate or SAC in the groups at baseline. Acute ingestion of flavanol-rich cocoa did not alter haemodynamics or SAC, nor was SAC affected by long-term ingestion of either dietary supplement (Figure 5) [SAC in the flavanol group−SAC in the placebo group=−0.010 (95% CI, −0.029 to +0.010), 0.017 (95% CI, −0.004 to +0.038), 0.006 (95% CI, −0.015 to +0.028) and −0.003 (95% CI, −0.025 to +0.018) at baseline, 90 min, 3 weeks and 6 weeks respectively].

Figure 5 SAC in patients ingesting non-flavanol (placebo)- or flavanol-containing cocoa at baseline, after 90 min (acute) and after 3 and 6 weeks

Biochemical parameters and circulating biomarkers of endothelial function

Biochemical parameters at baseline and at 6 weeks are shown for both flavanol and placebo (non-flavanol) groups in Table 4. There were no physiologically relevant differences in biochemical parameters measured at baseline and at 6 weeks in either group (flavanol or placebo). CRP, ICAM-1, E-selectin and P-selectin did not differ between the two groups at baseline or following 6 weeks of daily chocolate consumption. VCAM-1 levels were lower in the placebo (non-flavanol) group at baseline, but did not change appreciably with 6 weeks of treatment.

View this table:
Table 4 Biochemical parameters of the two groups

Values are presented as means±S.E.M., except CRP and lipoprotein (a) which are median values. *P=0.013, †P=0.005 and ‡P=0.010 when baseline values are compared with those in the placebo group.


In this randomized study, several well-established measurements of vascular endothelial function were examined to provide a comprehensive assessment of vascular function. The major finding was that flavanol-rich cocoa taken daily over a 6-week period was safe, but did not improve endothelial function or SAC in patients with multiple cardiovascular risk factors and advanced coronary atherosclerosis. Furthermore, flavanol-rich cocoa did not alter peripheral conduit vessel endothelial function 90 min after ingestion of a flavanol-rich cocoa beverage.

Flavonoids and endothelial function

In vitro studies have demonstrated that cocoa procyanidins and red wine flavonoids have endothelium-dependent vasorelaxant effects [9,11]. Potential mechanisms for these effects include the antioxidant properties of flavonoids; however, a more direct role via activation of eNOS (endothelial NO synthase) and increased NO production has been suggested [9,11,12]. Clinical studies examining the impact of short-term (2 week) purple grape juice consumption [25] and long-term (4 week) tea consumption [26] in patients with CAD have demonstrated that these flavanol-rich substances can significantly improve brachial artery FMD. In the context of these findings, the neutral effect of flavanol-rich cocoa on endothelial function in our present study is intriguing.

Conduit vessel endothelial function

In our study population, flavanol-rich cocoa or chocolate supplementation did not change conduit vessel arterial function in the acute or chronic setting. Baseline FMD and GTN responses were consistent in magnitude with the responses seen in other studies employing subjects of this age group and with this burden of vascular disease [17,25,27]. Our findings differ from the observations of Heiss et al. [8], who conducted a double-blind cross-over study of 20 subjects who were given cocoa and had indices of endothelial function measured 2 h later. In this acute study, ingestion of cocoa rich in flavanols (176 mg) resulted in an increase in brachial artery FMD from 3.4 to 6.3% (P<0.001), which was associated with surrogate evidence of increased NO bioactivity [8]. Intake of cocoa low in flavanols (<10 mg) did not result in any change in FMD. The subjects in the study by Heiss et al. [8] were on average 20 years younger and had fewer conventional cardiovascular risk factors than the subjects in our present study. These factors may have contributed to the differences observed.

Resistance vessel endothelial function

Examination of forearm resistance vessel function demonstrated that both endothelium-dependent and -independent responses were similar at baseline and were unchanged by 6 weeks of flavanol-rich cocoa consumption. There were no apparent differences in FBF responses to isotonic exercise and ischaemia, which are in part NO-dependent processes [21,22]. These findings concur with the lack of improvement in forearm reactive hyperaemia observed in the acute study by Heiss et al. [8], and suggest that resistance vessel endothelial function may not be improved by flavanol-rich cocoa.

Haemodynamics and arterial compliance

Despite previous suggestions that flavanol-rich chocolate might favourably affect BP in otherwise healthy subjects with mild untreated isolated systolic hypertension [28], we did not observe any significant treatment effect on systolic or diastolic BP, mean arterial pressure or heart rate. The use of concomitant vasoactive medication for the treatment of hypertension and ischaemic heart disease in the present study population, as detailed in Table 1, may have offset any potential antihypertensive effect of flavanol-rich cocoa.

Large artery stiffness is an independent predictor of future cardiovascular events in some patient groups with cardiovascular risk factors and is influenced in part by the endothelium [29]. Measurement of SAC provides an index of large proximal artery stiffness and may be improved by short-term dietary supplements [30]. Our findings of a lack of improvement in arterial compliance in subjects receiving flavanol-rich cocoa are consistent with the absence of change in endothelial vasodilator function.

Soluble biomarkers of endothelial function

Circulating plasma cellular adhesion molecules may represent surrogate markers for endothelial cell activation or damage [31]. We did not find any reduction in soluble ICAM-1, VCAM-1, E-selectin or P-selectin levels despite 6 weeks of flavanol-rich cocoa supplementation. Circulating oxidized LDL, a measure of oxidant stress, was also unchanged. These findings are in keeping with a study in healthy volunteers showing no effect of flavanol-rich cocoa (≈650 mg daily for 6 weeks) on markers of inflammation [IL-1 (interleukin-1), IL-6 (interleukin-6), TNF-α (tumour necrosis factor-α), hs-CRP (high-sensitivity CRP) and P-selectin] and urinary F2 isoprostanes [32].

Methodological issues

The two treatment arms were well matched, apart from a chance occurrence of a lower total cholesterol and LDL-cholesterol at baseline in the placebo group, indicating that the randomization process was effective. It is possible that the neutral result seen in the present study was due to the size of our study sample and that a smaller biological effect of flavanol-rich cocoa may not have been evident. Moreover, the lack of improvement on tests of vascular function with flavanol-rich cocoa does not signify a neutral impact on clinical vascular end points. Larger and longer-term studies will be required to address such clinical questions.

However, the sample size in our present study was similar to the study by Heiss et al. [8], in which positive effects on vascular function were observed. In contrast with previous work in the field, a strength of our present study was the use of multiple measures of vascular function. The lack of effect of flavanol-rich cocoa in improving any of these variables suggests that our findings are perhaps less likely to be a result of type 2 statistical error.

Inadequate flavanol intake cannot explain our findings. Compliance to the cocoa products was excellent and the total flavanol content within the cocoa products used in the present study (444 mg/day) was comparable with that used in other studies (176–821 mg/day) in which improvements of peripheral vascular function were noted [8,10,28]. It is possible that the age of subjects, and their burden of cardiovascular risk factors and vascular disease, may have been too great for flavanol-rich cocoa to exert a positive effect over the time frame of the study. Potential differences in baseline dietary flavonoid intake or the use of background medication (Table 1) in our subjects with chronic ischaemic heart disease, which are known to improve vascular function, may also have potentially masked any benefit of flavanol-rich cocoa. However, these medications were held constant during the study period. The findings of the present study do not support an incremental benefit of flavanol-rich cocoa on vascular function in subjects with CAD receiving typical therapies for this condition and its associated risk factors.


The present study demonstrates that consumption of flavanol-rich cocoa is not associated with an improvement of endothelial function in patients with multiple cardiovascular risk factors and CAD in the acute setting or over a 6-week period. Whether any benefit could be demonstrated in short- and long-term studies of younger subjects with single identifiable untreated cardiovascular risk factors should be the subject of future investigation.


Cocoa products and financial support for this study was provided by Mars Inc., Hackettstown, NJ, U.S.A. H.M.O.F. was supported by a Neil Hamilton Fairley Fellowship of the National Health and Medical Research Council of Australia. The authors have no financial or personal interests related to this study.

Abbreviations: ACh, acetylcholine chloride; BP, blood pressure; CAD, coronary artery disease; CI, confidence interval; CRP, C-reactive protein; FBF, forearm blood flow; FMD, flow-mediated dilation; GTN, glyceryl trinitrate; HDL, high-density lipoprotein; ICAM-1, intercellular cell-adhesion molecule-1; LDL, low-density lipoprotein; NO, nitric oxide; SAC, systemic arterial compliance; SNP, sodium nitroprusside; VCAM-1, vascular cell adhesion molecule-1; VOP, venous occlusion plethysmography


View Abstract