Chronic Consumption of Flavanol-rich Cocoa Improves Endothelial Function and Decreases Vascular Cell Adhesion Molecule in Hypercholesterolemic Postmenopausal Women

Abstract

Endothelial dysfunction characterizes many disease states including subclinical atherosclerosis. The consumption of flavanol-rich cocoa and cocoa-based products has been shown to improve endothelial function in both compromised and otherwise normal, healthy individuals when administered either acutely or over a period of several days, or weeks. Women experience increased risk for cardiovascular disease after menopause, which can be associated with endothelial dysfunction. Whether a flavanol-rich cocoa-based product can improve endothelial function in hypercholesterolemic postmenopausal women is not known. The purpose of the present study was to determine whether chronic dietary administration of flavanol-rich cocoa improves endothelial function and markers of cardiovascular health in hypercholesterolemic postmenopausal women. Thirty-two postmenopausal hypercholesterolemic women were randomly assigned to consume a high-flavanol cocoa beverage (high cocoa flavanols (CF)—446 mg of total flavanols), or a low-flavanol cocoa beverage (low CF—43 mg of total flavanols) for 6 weeks in a double-blind study (n=16 per group). Endothelial function was determined by brachial artery-reactive hyperemia. Plasma was analyzed for lipids (total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol), hormones (follicle-stimulating hormone), total nitrate/nitrite, activation of cellular adhesion markers (vascular cell adhesion molecule 1, intercellular adhesion molecule 1, E-Selectin, P-Selectin), and platelet function and reactivity. Changes in these plasma markers were then correlated to brachial reactivity. Brachial artery hyperemic blood flow increased significantly by 76% (P<0.05 vs. baseline) after the 6-week cocoa intervention in the high CF group, compared with 32% in the low CF cocoa group (P=ns vs. baseline). The 2.4-fold increase in hyperemic blood flow with high CF cocoa closely correlated (r²=0.8) with a significant decrease (11%) in plasma levels of soluble vascular cell adhesion molecule-1. Similar responses were not observed after chronic use of low CF. There were no significant differences between high and low CF in other biochemical markers and parameters measured. This study is the first to identify beneficial vascular effects of flavanol-rich cocoa consumption in hypercholesterolemic postmenopausal women. In addition, our results suggest that reductions in plasma soluble vascular cell adhesion molecule-1 after chronic consumption of a flavanol-rich cocoa may be mechanistically linked to improved vascular reactivity.

The vascular endothelium is the primary regulator of cardiovascular homeostasis in health and disease. It functions to maintain vascular tone, regulate permeability, attenuate vascular inflammation, and inhibit smooth muscle cell proliferation.¹ Early atherosclerosis is characterized by a number of processes that include prelesional endothelial dysfunction, endothelial activation, and monocyte recruitment to the endothelium and subendothelial space. Endothelial dysfunction and impairment in endothelial-dependent vasodilation characterize many disease states, including cardiovascular disease (CVD).^2,3 Common contributors to endothelial dysfunction include cardiovascular risk factors such as elevated nascent and modified low-density lipoprotein cholesterol (LDL-C), smoking, hypertension, diabetes mellitus,³ and the menopausal state.⁴ Activation of the endothelium is typically associated with increased concentrations of soluble adhesion molecules in the blood, including intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1), P-Selectin, and E-Selectin² that can participate in the recruitment of monocytes and other blood-borne cells.

Clinical studies indicate that high dietary intake of the antioxidant vitamins, folic acid, L-arginine, and ω-3 fatty acids favorably affects vascular endothelial function.^2,5,6 Plant-derived foods rich in polyphenolic compounds, such as the flavonoids, have also received a great deal of attention for their potential cardiovascular health benefits, and certain cocoas can be an especially rich source of a specific subclass of these phytochemicals known as flavanols. Cocoa flavanols (CFs) and flavanol-rich cocoa products have demonstrated potential, both in vitro and in human studies, to modulate cardiovascular health in at least 4 important ways, including: improved endothelial function, modulation of inflammation, reduced platelet reactivity, and increases in oxidative defense. For example, the consumption of flavanol-rich cocoa and cocoa-based products increases plasma antioxidant capacity and lag time to LDL oxidation.^7–13 In addition, the consumption of a flavanol-rich cocoa or cocoa-based product has been shown to reduce platelet reactivity, as determined by assays measuring platelet aggregation and primary hemostasis, and marker of platelet activation.^11,14–17 Consumption of flavanol-rich cocoa beverages has been shown to improve endothelial function in healthy smoking and nonsmoking subjects.^18–23 Finally, although the duration of the effect and the mechanism(s) of action are apparently quite different, the effect of flavanol-rich cocoa on platelet activation is qualitatively similar to that after aspirin administration on an acute basis, albeit of lesser magnitude.^16,24

Flavonoid-rich beverages favorably affect endothelial function in healthy men and women subjects, and those with vascular compromise. For example, vascular endothelial function in subjects with established CVD improved after the consumption of flavonoid-rich beverages.^5,25,26 Similarly, in subjects with no history of CVD, flow-mediated dilation (FMD) after the consumption of dealcoholized red wine²⁷ and black tea²⁸ increased significantly (5.6% and 2.1%, respectively). Heiss et al²² demonstrated that flavanol-rich, but not flavanol-poor, cocoa could significantly increase both FMD and the nitric oxide (NO) pool 2 hours after ingestion of a single dose (176 mg of total flavanols). They have recently shown that flavanol-rich cocoa (176 to 185 mg/dose) can reverse endothelial dysfunction in a group of smokers, also at 2 hours postconsumption.²¹ Importantly, Fisher et al¹⁹ added substantially to the field by unambiguously demonstrating that the enhancement in peripheral vasodilation in otherwise normal, healthy subjects after ingestion of flavanol-rich cocoa (821 mg of total flavanols per day) was NO-dependent.

Although the literature describing flavanol-rich cocoa and its ability to improve important markers of cardiovascular health is impressive, additional questions remain to be answered. It is unknown, for example, whether sex differences exist with respect to endothelial response to flavanol-rich cocoa. It is also not known whether premenopausal and postmenopausal women differ in their endothelial response to flavanol-rich cocoa, and how this may be modulated by hyperlipidemia. Thus, in the present study, we investigate the effects of flavanol-rich cocoa consumption on markers of cardiovascular risk and endothelial function in hypercholesterolemic, postmenopausal women, not taking hormone-replacement therapy (HRT) or cholesterol lowering medication. As this group represents individuals in the population who are at risk for subclinical atherosclerosis because of their age, menopausal status, and hyperlipidemia,^4,29–31 we hypothesized that they would display a cardioprotective response after the intake of flavanol-rich cocoa. In addition, because little is known regarding whether the reported vascular actions of acute flavanol-rich cocoa administration can persist during chronic use, we also chose to investigate the impact of chronic flavanol-rich cocoa consumption by postmenopausal women.

METHODS

Subjects

The study group consisted of 32 hypercholesterolemic postmenopausal women. Subjects were nonsmokers, did not use cholesterol-lowering drugs, and did not regularly use antioxidant supplements. Participants had no history of CVD, thyroid disorder, or diabetes mellitus. Inclusion criteria included a fasting serum total cholesterol (TC) level >200 mg/dL, postmenopausal state (defined by absence of a menstrual period for 12 consecutive months and follicle-stimulating hormone (FSH) level of 23 to 116.3 mIU/mL), body mass index<30 kg/m², and no use of HRT. Those previously on HRT were asked to stop HRT with the consent of their physician, and washout for 6 weeks before starting the study. Study participants provided written, informed consent, following protocols approved by the Institutional Review Board at the University of California, Davis.

Study Design

The study was a randomized, double-blind, parallel-arm study with 2 treatment groups: a high-flavanol cocoa beverage (high CF), and a low-flavanol cocoa beverage (low CF) (n=16 women in each group). Subjects were entered on a rolling-enrollment basis, and the study was conducted over a total of 9 months. All subjects participated in the 11-week study protocol, and made up to 6 visits to the clinic (Fig. 1). Low CF was consumed daily for the first 2 weeks of the study (week 0 to 2, run-in), subjects were then randomized to consume high or low CF daily for a 6-week period (week 2 to 8, treatment period), followed by a 3-week washout period (week 8 to 11) during which time no cocoa beverage was consumed.

FIGURE 1.: Schematic summary of study design and timeline of clinic visits and blood draws. BD, blood draw; BR, brachial artery reactivity; CF, cocoa flavanols.

A 3-day diet record was kept at 4 time points throughout the study, at weeks 0, 2, 8, and 11. Subjects followed a low-flavonoid diet for 24 hours before each clinic visit by avoiding all fruits and vegetables, and juices derived from them, wine, and tea. Cocoa and other chocolate products were avoided during the entire 11-week study period. Subjects otherwise maintained their regular diet.

Cocoa Beverage

Subjects were supplied with individually sealed servings of 36 g of cocoa powder (Mars, Incorporated, Hackettstown, NJ) labeled with a 3-digit code, and asked to reconstitute it in 240 mL of hot water. To assess compliance, empty packets and any unused servings were collected at follow-up visits and counted. The cocoa powder consisted of 18.8 g cocoa powder, 18 g sucrose, and trace amounts of carrageenan, vanillin, cinnamon, and salt (for suspension of solids and flavoring). High CF contained an average of 12.4 mg total CFs/g and 446 mg of total flavanols. Low CF beverage contained an average of 1.2 mg total CFs/g and 43 mg of total flavanols.

Sample Collection

Fasting (12 h) blood samples were obtained from study participants at weeks 0 (baseline), 2, 8, and 11. Subjects were instructed to refrain from the use of nonsteroidal anti-inflammatory medications for 4 days before a blood draw. To assess changes in markers of vascular function, lipids, and platelet reactivity, blood was collected in evacuated tubes containing ethylene diaminetetraacetic acid, sodium heparin, or sodium citrate, respectively. Plasma was separated by low-speed centrifugation (1800g for 15 min at 4°C) and stored at −80°C until the end of the study. Platelet reactivity was determined on whole blood within 4 hours of collection using a Platelet Function Analyzer (PFA-100; Dade Behring International, Miami, FL). For a given assay, individual samples from all of the study time points were analyzed concurrently to prevent interassay variation. Blood was also collected for fasting TC, LDL (LDL-C) and high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), and glucose. Lipid and chemistry panels were performed at the UC Davis Medical Center clinical pathology laboratory using a Synchron LX-20 System (Beckman Coulter Inc, Brea, CA). LDL-C was calculated using the Friedewald equation.³² Postmenopausal status of study participants was confirmed by assay of serum FSH at the UC Davis Medical Center clinical lab using a 2-site chemiluminotic immunoassay (Bayer Diagnostics, Medfield, MA). Subject's height was measured once at the first visit, and at this and subsequent visits weight, blood pressure, and heart rate were recorded.

Plasma Nitrate and Nitrite

Nitrate and nitrite levels were measured in both plasma and urine samples by a modified method of the Griess reaction that converts all nitrate to nitrite using the bacterial enzyme nitrate reductase.³³ The colorimetric reaction between nitrite, sulfanilamide and N-(1-naphthyl) ethylenediamine produces a pink/magenta azo product with a maximum absorbance at 543 nm. Absorbance was read using a Multiskan Ascent microplate photometer (Labsystems Inc, Franklin, MA).

Biochemical Markers of Endothelial Function and Cardiovascular Risk

To determine whether changes in brachial artery reactivity were associated with changes in biochemical markers of endothelial function and vascular inflammation, biomarkers were measured on plasma samples collected at weeks 2, 8, and 11 using standard commercially available enzyme immunoassay kits following the manufacturer's instructions. Markers included those associated with early monocyte binding to endothelium and endothelial activation including: soluble intercellular adhesion molecule-1 (sICAM-1), soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble P-Selectin (sP-Selectin), and soluble E-Selectin (sE-Selectin).

Assay of Platelet Function

To determine if chronic consumption of flavanol-rich cocoa affects platelet reactivity, platelet function was measured with whole citrated blood using the Platelet Function Analyzer (PFA-100) (Dade Behring International, Miami, FL) as described previously.¹⁶ Care was taken to ensure that blood collected for this assay remained undisturbed for at least 30 minutes before analysis, and only blood drawn within 4 hours before performing the assay was used. The PFA-100 measures collagen-epinephrine or collagen-ADP-induced hemostatic plug formation under simulated small vessel shear conditions (5000 to 6000/s), and the time in seconds to occlude the aperture of a collagen-coated membrane (closure time). If occlusion of the aperture did not occur by 300 seconds, the test was stopped. Samples were processed in duplicate.

Determination of Plasma Epicatechin

Heparinized plasma samples were extracted and analyzed for plasma epicatechin concentration. For these studies, an HP 1100 high-pressure liquid chromatography system with Chemstation software, equipped with a quaternary pump, temperature-controlled autosampler, column oven, and diode array detector (Hewlett-Packard, Wilmington, DE) in series with a CoulArray 5600 detector (ESA, Chelmsford, MA) was used as previously described.³⁴

Brachial Artery Reactivity

FMD of the brachial artery was measured at the beginning and end of the 6-week cocoa intervention in each study subject. Changes in vessel diameter after reactive hyperemia (endothelium-dependent vasodilation), and after sublingual nitroglycerin (endothelium-independent vasodilation) were measured according to previously described methods.³⁵ Brachial artery ultrasound scans were performed after an overnight fast with the subjects at rest after lying in the supine position for at least 10 minutes. Studies were performed using a Phillips HDI 3500 ultrasound machine and a high-resolution (7.5 MHz) linear array transducer by 2 experienced vascular ultrasound technicians. The variability in measures between the 2 technicians was <5%.

The right brachial artery was imaged above (3 to 7 cm) the antecubital fossa and arterial bifurcation with a vascular probe positioned at an angle of 60 degrees. First, baseline brachial artery diameter and flow were measured. Next, a blood pressure cuff was inflated to suprasystolic pressure (at least 30 mm Hg higher than the subject's systolic blood pressure) and the vessel compressed for 5 minutes. The cuff was deflated and brachial artery diameter and flow were measured within 15 seconds after decompression during reactive hyperemia to yield the postocclusion data. A second baseline brachial artery diameter and flow were measured after 10 minutes to determine that the vessel returned to baseline. Then, nitroglycerin was administered sublingually as a tablet (0.4 mg) and the postnitroglycerin brachial artery flow velocity and diameter were determined after 4 minutes.

All subjects participated in the vascular function protocol. The following parameters were measured during each ultrasound session at baseline, postocclusion, at a second baseline, and after nitroglycerin administration: systolic peak flow velocity (cm/s), vessel diameter (mm), mean flow velocity (cm/s), and flow volume by diameter (cm/s). FMD was calculated as the percent change in vessel diameter postocclusion compared with the first baseline. Hyperemic flow was defined as blood flow in mL/min during reactive hyperemia and calculated as postocclusion systolic peak flow velocity×(π)×(radius of brachial artery)² and reported as mL/min. Change in peak blood flow was calculated as the percent change in hyperemic flow postocclusion compared with the baseline value. To calculate endothelium-independent vasodilation, nitroglycerin-mediated dilation was determined using the postnitroglycerin arterial diameter value compared with the second baseline.

Statistical Analysis

Data are presented as the mean±SEM. Statistical analysis was performed using SigmaStat for Windows version 2.03 (SPSS Inc) and SAS for Windows version 8.1 (SAS Institute, Cary, NC). Data were analyzed by 2-way repeated measures analysis of variance (ANOVA). General linear models were used to examine differences in postintervention values compared with baseline values. Logarithmic transformations were performed on all variables. Differences were considered significant at P≤0.05. All parameters were correlated using the Pearson product correlation.

RESULTS

Subject Characteristics

Baseline subject characteristics were similar between the 2 study groups (high and low CF), and there were no statistically significant differences between groups in any of the demographic variables we determined. Mean age of the subjects in the high and low CF groups was 57.7±2.2 and 55.4±1.7 years, respectively. The high and low CF groups had an average body mass index of 24.9±1.0 and 25.3±0.8 kg/m², respectively. There were no differences in FSH levels between the high and low CF groups (90.7±6.6, 81±7.4, respectively), and all women were confirmed to be postmenopausal (FSH levels 23 to 116.3 mIU/mL). Three women in the high CF group, and 2 women in the low CF group were on hormone-HRT before the study, and with the consent of their physician, stopped using HRT 6 weeks before entering the study. Subjects that were subsequently determined to have abnormal vascular responses, defined by paradoxical vasoconstrictor responses to vessel occlusion or no change in vascular reactivity in response to vessel occlusion (a total of 15 women), were excluded from subsequent statistical analysis as these findings suggested underlying vascular dysfunction and possible subclinical vascular disease. This heterogeneous subset includes: 5 subjects (3 high CF and 2 low CF) that were unable to provide a FMD measurement at week 8, 4 subjects (1 high CF and 3 low CF) that had a negative FMD percent change at week 2, 1 high CF subject with no percent change in FMD at week 2, and 5 subjects (2 high CF and 3 low CF) with negative FMD percent changes at week 8. Thus, data from a total of 9 subjects (56% of women) in the high CF group and 8 subjects (50% of women) in the low CF groups were used for the lipid, biochemical, and vascular function analyses.

Dietary Analysis

Individual food intake records from weeks 0, 2, 8, and 11 were analyzed using Nutritionist V (First Data Bank, San Bruno, CA), and data from 3-day diet records at each time point averaged. As shown in Table 1, dietary intake of energy, macronutrients, fats, vitamins, and other substances did not change significantly over the 6-week treatment period in either cocoa treatment study group. In addition, no significant differences were observed in any of the dietary parameters between the 2 cocoa treatment groups.

TABLE 1: Dietary Nutrient Intake for Subjects in Each Treatment Group During the 6-week Cocoa Beverage Intervention Period

Hemodynamic Parameters

There were no significant changes observed in systolic or diastolic blood pressure in the high CF group over the 6-week dietary intervention period, Table 2. In contrast, in the low CP group, both systolic and diastolic blood pressure decreased significantly by 9.3% and 6.5%, respectively, after the 6-week dietary intervention period (P<0.05). In addition, diastolic blood pressure values at week 2 were also significantly (P<0.01) different in the high and low CF groups. There were no significant changes in heart rate in either group after the cocoa beverage interventions.

TABLE 2: Brachial Artery Reactivity and Hemodynamic Parameters

Brachial Artery Reactivity

There were no differences in resting or postnitroglycerin arterial diameter at the beginning and end of the intervention in either of the CF dietary groups. As expected, hyperemic diameter and postnitroglycerin diameter increased from baseline in both CF beverage study groups. After 6-week CF dietary intervention, FMD increased 2% in the high CF group, and decreased 1.5% in the low CF group. The increase in FMD in the high CF displayed a strong trend toward statistical significance (P=0.0585). Furthermore, in the high CF beverage group hyperemic blood flow increased by 76% compared with baseline (P=0.002), consistent with a significant vasodilatory effect. Compared with baseline, hyperemic blood flow increased by 32% in the low CF group; however, this change did not reach statistical significance (Fig. 2).

FIGURE 2.: Change in hyperemic blood flow from week 2 to 8 for the high and low CF study groups. The increase was significantly different in the high CF cocoa beverage group, P=0.002 [all pairwise multiple comparison procedures (Tukey test)].

Adhesion Markers

Levels of biochemical markers of endothelial function and adhesion molecules during the 6 weeks on the cocoa treatments are reported in Table 3. After 6-week intervention, sVCAM-1 decreased by 10.7% in the high CF group (P=0.009), suggesting a significant attenuation of activation of this marker of vascular health in response to daily flavanol-rich cocoa intake. In contrast, sVCAM-1 increased in the low CF group. The differential response between high and low CF on sVCAM-1 was highly statistically significant (P=0.010). Because considerable attention has focused on soluble adhesion molecules as early biomarkers of alterations in vascular function, we investigated whether in the high CF study group there was a relationship between the decrease in sVCAM-1 levels and the increase in hyperemic flow. This analysis demonstrated these parameters to be tightly correlated (r=0.814, P=0.014) in a unique way as there were no significant changes in sP-selectin, sE-selectin, or sICAM-1 over the 6-week study period for either cocoa treatment group. In addition, there were no significant differences in these parameters, or total nitrates and nitrites, between the 2 cocoa treatment groups.

TABLE 3: Biochemical Markers of Endothelial Function

Plasma Lipids

The mean starting serum lipid concentrations for the high and low CF groups, respectively, were: 235±8 versus 240±8 mg/dL (TC), 149±5 versus 150±9 mg/dL (LDL), 69.7±4.6 versus 68.1±6.1 mg/dL (HDL-C), 3.5±0.2 versus 3.9±0.3 (TC:HDL ratio), and 79.7±7.8 versus 111.1±21.1 mg/dL (TGs). Although TC, LDL-C, and TG did not change within either treatment group after 6-week intervention, there was a differential effect of cocoa dose on HDL-C. HDL-C levels increased by 6.6% with high CF, yet decreased by 9.6% in the low CF group. This differential response on HDL-C was statistically significant, P<0.05. However, changes in HDL-C did not correlate with changes in hyperemic flow or any of the other cardiovascular health markers studied (data not shown). Measurement of blood lipid concentrations at week 11 indicated no significant changes to TC, LDL, the TC:HDL ratio, and TG. The significant difference in the concentration of serum HDL-C between the groups remained at week 11.

Platelet Function and Flavanol Consumption

After 6 weeks of the daily cocoa beverage intervention, there was a 13% increase in ADP/collagen-induced PFA-100 closure time in the high CF compared with the low CF group that approached statistical significance (P=0.079). Baseline values for ADP/collagen-induced platelet reactivity in the high and low CF groups were 84.5±2.2 and 96.6±8.9 seconds, respectively. There were no significant changes in epinephrine/collagen-induced platelet reactivity in either CF treatment group after 6-week dietary intervention. Baseline values for epinephrine/collagen-induced platelet reactivity were 169.1±17.5 and 173.2±22.9 seconds for the high and low CF groups, respectively.

DISCUSSION

The objective of this dietary intervention study was to determine whether chronic (6 weeks) consumption of flavanol-rich cocoa attenuated cardiovascular risk markers and improved vascular function, as measured by FMD and reactive hyperemia of the brachial artery, in hypercholesterolemic postmenopausal women. Our findings extend prior observations of improved markers of cardiovascular health resulting from consumption of flavanol-rich cocoa either acutely or over a period of several days. In addition, a vasculoprotective effect for chronic flavanol-rich cocoa was demonstrated in hypercholesterolemic postmenopausal women, a population not previously studied with regard to effects of flavanol-rich cocoa consumption. Our findings can be summarized as follows: (1) chronic consumption of high CF, but not low CF, improves FMD and significantly increases hyperemic brachial artery blood flow; (2) the improvement in endothelial function with high CF seems to be mechanistically linked to reductions in sVCAM-1; and (3) chronic consumption of high CF but not low CF leads to a significant increase in HDL cholesterol.

The endothelium is important for maintaining normal vascular function and preventing early events in atherogenesis,^3,36,37 and endothelial dysfunction can precede appreciable atherosclerotic lesion development. Noninvasive imaging with Doppler ultrasound³⁸ has revealed changes in endothelium-dependent FMD in patients with risk factors for vascular disease. Improvement in FMD has been previously demonstrated with a number of dietary interventions, including soy protein³⁹, and cocoa.^18,21–23 To our knowledge, this study is the first to examine the effect of chronic consumption of a flavanol-rich food on endothelial function and vascular biomarkers in mildly hypercholesterolemic postmenopausal women. Our study used a number of unique features compared with prior work. Previous studies of flavanol-rich cocoa have used male subjects, or mixed study groups of both males and females. In addition, although the menopausal status of women has not been previously noted, on the basis of the age of subjects it seems that prior investigations have not included postmenopausal women.^18–23 Postmenopausal women are at a greater risk for CVD because of their age and menopausal status, and as our findings demonstrate, this is a group of individuals that can exhibit early impairment of vascular function. Therefore, the present study focused on postmenopausal women to determine whether flavanol-rich cocoa could influence risk markers associated with CVD. Furthermore, our study population was mildly hypercholesterolemic, permitting us to ascertain whether flavanol-rich cocoa can favorably impact lipid parameters. Lastly, we also measured changes in vascular reactivity in response to chronic cocoa consumption (6 weeks of daily flavanol-rich cocoa consumption) to compare the vascular responses to the acute or shorter duration administration of flavanol-rich cocoa used in prior studies.^18–23 Because the effect of chronic consumption of flavanol-rich cocoa has not been well studied and there are numerous examples of acute drug effects differing from chronic responses, we sought to provide longer duration consumption in our study design.

After 6-week high CF dietary intervention, we observed a significant increase in hyperemic blood flow, an important marker of vasodilation, and a 2% increase in FMD. Although the change in FMD approached statistical significance it did not reach it. However, the increase approximated the 3% increase in FMD noted by Heiss et al²² after acute (2 h) consumption of a flavanol-rich cocoa drink and a 1.3% increase after dark chocolate consumption,¹⁸ and a 2.1% increase in FMD after black tea consumption as reported by Hodgson et al²⁸ It is important to note that the FMD measurements in these studies were all taken within hours after consumption of the flavonoid-rich food. Short-term flavanol-rich cocoa supplementation studies suggest that plasma epicatechin partially contributes to an improved FMD by increasing NO availability, a response that peaks ∼2 hours after cocoa consumption.^{11,19,21,40–42} However, we were unable to detect significant levels of nonmethylated plasma epicatechin after the overnight fast that preceded the FMD measurements. If the positive effects of cocoa supplementation are transient in nature, dissipating within 1 to 2 hours after stopping cocoa consumption, this could in part explain our results. However, little is known about the long-term effects of flavanol consumption, and as recently observed for quercetin,⁴³ it may be possible for flavanols to accumulate within tissues over a more prolonged period of time, to produce a biologic effect. Therefore, it is important to note that all subjects were given the low-flavanol cocoa for a 2-week run in period before the beginning of the trial. This “low-flavanol” cocoa was not completely devoid in epicatechin or catechin content, but rather contained 43 mg of total flavanols. A 2-week supplementation study providing 46 mg/day of epicatechin in dark chocolate produced significant improvements in FMD 2 hours after the last chocolate dose.¹⁸ Recently, Heiss et al,²¹ reported improvements in FMD response with a drink containing ∼70 mg of epicatechin and catechin. In addition, we have reported reduced platelet reactivity with as little as 220 mg of total flavanols (in a small handful of chocolate chips).¹¹ Epidemiology studies also indicate a 25% lowered risk in ischemic heart disease individuals that consume 50 mg of catechin in the diet per day.⁴⁴ Thus, it is possible that the 2-week run-in period with the low flavanol cocoa beverage improved the FMD baseline to such an extent that the addition of the high flavonoid drink did not provide statistically significant improvements. The above issue should be addressed in future studies.

To determine the potential mechanism underlying the increase in hyperemic flow observed with high CF in our study, we investigated a number of important variables previously demonstrated to affect vascular function. For this analysis we considered NO, serum lipids, platelet aggregation, and vascular cell adhesion markers. Our findings with each of these parameters are discussed below.

We first considered the potential role of NO. It has been suggested that changes in NO signaling after consumption of flavanol-rich foods may be responsible for their vasodilatory effects. Indeed, the administration of the NO synthase inhibitor, L-NAME, completely prevented the dilatory effects induced by cocoa consumption in studies by Fisher et al.¹⁹ Although we observed no differences in plasma total nitrite and nitrate after consumption of the high or low CF cocoa beverage for 6 weeks, this does not preclude the possible involvement of early effects because of an increase in NO availability. As total plasma nitrate and nitrite was measured after an overnight fast, we may have failed to detect any short-term transient effects of cocoa that could have occurred shortly after the consumption of the cocoa beverages.

Higher plasma HDL-C is atheroprotective,⁴⁵ and it is associated with improved endothelial function in hypercholesterolemic patients.⁴⁶ In addition, elevation of HDL-C can restore endothelial function in patients with low HDL-C.⁴⁷ We did not observe a significant difference in HDL-C during the 6-week dietary cocoa intervention in either flavanol-rich cocoa treatment group. However, compared with baseline values, the increase in HDL-C observed with high CF was significantly different from the decrease in HDL-C observed in the low CF group. The change in HDL-C with flavanol-rich cocoa dose observed in the current study may be important, particularly given the fact that a similar observation was made by Wan et al.⁹ These results suggest that an improved lipid biomarker profile could be mechanistically linked to the observed improved vascular responsiveness after chronic high-dose cocoa in postmenopausal women.

Next, we considered the potential role of soluble adhesion molecules. The expression, principally on endothelial cells, of molecules such as VCAM-1, ICAM-1, and E-selectin mediates leukocyte recruitment to sites of vascular inflammation and is an early event in atherogenesis.^48,49 In animal models, the expression of VCAM-1 and ICAM-1 localizes to areas predisposed to the formation of atherosclerotic plaque.^48,50 Thus, considerable attention has focused on soluble adhesion molecules as early biomarkers of alterations in vascular function, as they are indirect measures of vascular inflammation and endothelial cell activation. In animal models of atherosclerosis, for example, the use of antibodies to adhesion markers has been shown to reduce various CVD parameters.⁵¹ Evidence that the concentration of soluble adhesion markers (eg, sVCAM-1 and sICAM-1) in plasma are predictive of cardiovascular risk has led to the suggestion that they could be targets for therapeutic intervention.⁵¹ Thus, identifying a food product, or compounds contained in food (such as flavanols), as potential modulators of adhesion molecule expression is of considerable interest. We observed that subjects who consumed the high CF had significantly lower levels of sVCAM-1 after 6-week intervention, compared with those consuming the low CF beverage. Furthermore, the increase in hyperemic flow strongly correlated with the decrease in sVCAM-1 in this same group of postmenopausal women. This suggests that the chronic use of high flavanol cocoa in postmenopausal women can improve endothelial function; on the basis of our findings, this may involve a VCAM-1-dependent mechanism.

It is not yet clear which specific components in the flavanol-rich cocoa provided to subjects in our study contributed to the observed improved vascular reactivity response and concomitant reduction in VCAM-1. Similar to the cocoa used by Fisher et al,¹⁹ ingredients in both the high and low CF used in this investigation were similar in composition, except for the amount of total flavanols. However, given that epicatechin and certain B-type dimer(s), and their related metabolites are the presumed predominant flavanols in human plasma after cocoa consumption,^34,52 they are candidates of potential interest for the effects of high CF observed in this and other studies. These flavonoids are reported to inhibit activation of the oxidative-stress sensitive nuclear transcription factor, nuclear factor kappa-B (NF-κB). It is pertinent to note that in so doing flavanols can inhibit the upregulation of VCAM-1 endothelial expression in vitro.^53,54 As the transcription factor NF-κB is a known promoter of VCAM-1 expression,³⁹ it can be speculated that reductions in NF-κB activation in cells of the vascular endothelium provide a mechanistic explanation for a flavanol-induced decrease in sVCAM-1. Future studies are needed to investigate the downstream signaling events associated with of flavanol-rich cocoa-induced decreases in VCAM-1.

Certain flavonoids may be partly responsible for a decreased tendency for platelet activation and aggregation, important mechanisms involved in the progression and pathogenesis of CVD. We have also previously demonstrated in humans that an acute dose of a flavanol-rich cocoa product inhibited platelet activation and function as evidenced by changes in epinephrine/collagen-induced platelet reactivity.¹⁶ We have also previously demonstrated that an acute dose of flavanol-rich cocoa inhibits platelet activation and function over 6 hours in healthy humans.¹⁵ The mechanism, in part, is via suppression of ADP-stimulated or epinephrine-stimulated platelet activation and platelet microparticle formation, leading us to conclude that flavanol-rich cocoa consumption can have an acute effect on primary hemostasis that is similar to aspirin. However, we have not yet studied the effects of chronic consumption of flavanol-rich cocoa, nor have we studied this in postmenopausal women. Therefore, the focus of the platelet-related protocols of this study was to investigate the longer-term effect of high and low CF on platelet reactivity in this group. We observed a trend toward an increase in ADP-stimulated platelet aperture closure time after consumption of the high CF product. Our findings are in agreement with other investigators who demonstrated that flavanol-rich cocoa supplementation for 28 days significantly increased plasma epicatechin and catechin concentrations and significantly decreased platelet function.⁵⁵

The consumption of a heart-healthy diet as an adjunct to therapeutic lifestyle changes is widely recommended for cardiovascular risk reduction.²⁹ Although the total number of subjects included in our statistical analyses was constrained by the fact that a significant number of subjects exhibited abnormal baseline vascular function our findings suggest that in hypercholesterolemic postmenopausal women daily consumption of flavanol-rich cocoa can significantly improve vascular endothelial function. In addition, the vascular improvements are tightly correlated to a reduction in sVCAM-1 and associated with a increase in HDL-C, all known predictors of CVD risk. The results from this present study indicate that the chronic consumption of flavanol-rich cocoa can be beneficial for hypercholesterolemic postmenopausal women, extending the growing body of evidence that flavanol-rich diets provide significant cardiovascular protection. Additional studies investigating the actions of other dietary flavonoids having putative cardiovascular benefits in postmenopausal women could therefore be of interest.

ACKNOWLEDGMENTS

The authors thank Francine Nardinelli, Deborah Finley, and Holly Beals for providing technical assistance with the ultrasound exams. They also thank Neil Willits for providing the statistical analysis for this study. J.F.W.P., H.H.S., and C.L.K. contributed to the concept of the study. J.F.W.P., A.C.V., J.A.P., H.H.S., and C.L.K. contributed to the design of the study. J.F.W.P., J.A.P., L.L., R.R.H., H.S., J.L.E., and F.M.S. contributed to the conduct of the study. J.F.W.P., A.C.V., J.A.P., and C.L.K. contributed to the interpretation of the data. Sherri Lazarus assisted the investigators in obtaining the cocoa product.