|Realistic intake of a flavanol-rich soluble cocoa product increases HDL-cholesterol without inducing anthropometric changes in healthy and moderately hypercholesterolemic subjects
Sara Martínez-Lópeza&, Beatriz Sarriáa*&, José Luis Sierra-Cinosb, Luis Goyaa, Raquel Mateosa, Laura Bravoa
aDepartment of Metabolism and Nutrition. Institute of Food Science, Technology and Nutrition (ICTAN-CSIC) Consejo Superior de Investigaciones Científicas (CSIC), José Antonio Nováis 10, 28040 Madrid, Spain.
bDepartment of Nutrition and Bromatology I. School of Pharmacy. Complutense University of Madrid (UCM). Ciudad Universitaria, s/n. 28040 Madrid, Spain.
Address: Institute of Food Science, Technology and Nutrition (ICTAN-CSIC), José Antonio Novais 10, 28040 Madrid, Spain.
Tel: +34915492300, FAX: +34915493627,
& Sara Martínez-López and Beatriz Sarriá have contributed equally to this work.
Running title: Cocoa enhances HDL-C without anthropometric changes
Key words: cocoa, polyphenols, flavanols, HDL-cholesterol, cytokines, adhesion molecules, antioxidant, inflammation, humans, hypercholesterolemia
To assess whether antioxidant, anti-inflammatory and other cardio-protective effects attributed to cocoa are achieved when regularly consuming moderate amounts of a flavanol-rich soluble cocoa product, a non-randomized, controlled, crossover, free-living study was carried out in healthy (n=24; 25.9±5.6y) and moderately hypercholesterolemic (200-240 mg/dL; n=20; 30.0±10.3 y) volunteers. Participants consumed two servings/day (7.5 g/serving) of a soluble cocoa product (providing 43.05 mg flavanols/day) in milk, which was compared with consuming only milk during a 4 week period. The effects on systolic and diastolic blood pressure and heart rate were determined, as well as on serum lipid and lipoprotein profiles, interleukins (IL)-1β, IL-6, IL-8, IL-10, tumor necrosis factor-α (TNF-α), monocyte chemoattractant protein-1 (MCP-1), vascular (VCAM-1) and intercellular cell adhesion molecules (ICAM-1), serum malondialdehyde (MDA), carbonyl groups (CG), ferric reducing/antioxidant power (FRAP), oxygen radical absorbance capacity (ORAC), and free radical scavenging capacity (ABTS). During the study, the volunteers’ diets and physical activity were also evaluated, as well as any changes in weight, skin folds, circumferences and related anthropometric parameters. Cocoa and certain polyphenol-rich fruits and vegetables and their derivatives were restricted. After consuming the cocoa product positive effects were observed such as an increase in serum HDL-C (p<0.001) and dietary fiber intake (p=0.050), whereas IL-10 decreased (p=0.022). Other cardiovascular-related biomarkers and anthropometric parameters were unaffected. We have therefore concluded that regular consumption of this cocoa product in a Spanish-Mediterranean diet may protect against cardiovascular disease in healthy and hypercholesterolemic subjects without producing any weight gain or other anthropometric changes.
A diet rich in flavonoids promotes health and delays the onset of cardiovascular disease. Cocoa contains higher concentrations of flavonoids per serving than tea or red wine,1 and has been defined as a functional food due to its high flavanol content.2 Even so, although flavanols in cocoa are among the most powerful antioxidants identified so far, their bioavailability is relatively low and depends on the matrix in which the polyphenols are delivered.3,4 Proteins of the food matrix in which cocoa is consumed might bind flavonoids, reduce their bioavailability and consequently their potential antioxidant properties in vivo.5 Recently, polyphenols contained in a water-insoluble cocoa fraction showed antioxidant action despite being bound to macromolecules using an in vitro model,6 and there are also in vivo reports on the antioxidant potential of cocoa flavanols in spite of their limited bioavailability.5 After the acute intake of cocoa products an increase in plasma antioxidant status7 and a decrease in oxidation parameters8 have been described. However, these effects could not be sustained after repeated cocoa consumption. In other chronic studies, varied effects were observed such as a reduction in MDA levels9 and plasma-oxidized LDL concentrations;10 however, F2-isoprostane levels remained unchanged in both plasma11 and urine.12
Apart from its antioxidant properties, cocoa has also been shown to have a beneficial effect on endothelial function, lipemia, inflammation and blood pressure, as well as on other cardiovascular-protective properties, which have been attributed to its polyphenol fraction.2,4 Nevertheless, to date a cause-effect relationship between the consumption of cocoa flavanols and decreased lipid oxidation has not yet been unequivocally established.13,14 Conversely, there is enough evidence to establish a relationship between cocoa flavanols and maintenance of endothelium-dependent vasodilation, which contributes to normal blood flow.15
Most of the knowledge on cocoa flavanols and inflammation has been obtained using in vitro experimental models.2 More in vivo studies are necessary to understand the cumulative effects of regularly consuming cocoa products on cell adhesion molecules and inflammatory biomarkers, since these effects seem to depend on the amount of cocoa consumed16 and its flavanol content.17 Moderate consumption of cocoa products inhibits platelet activation and aggregation in healthy and cardiovascular disease risk subjects.4,18
Although cocoa products are high-energy foods, they have been shown to have anti-obesity effects in humans5 and rats.19 In addition, chocolate and other cocoa soluble product manufacturers are actively pursuing ways of producing novel low energy products by lowering sugar and fat levels without compromising the flavor and texture of traditional chocolate.
In view of the foregoing, this work questions whether cardiovascular-related effects can be obtained by regularly consuming a moderate and realistic amount of cocoa powder. The present work has evaluated the effects of regularly consuming a moderate amount of a novel cocoa powder, rich in flavanols and low in sugar, within a balanced Mediterranean-Spanish diet, on cardiovascular related and parameters in healthy and moderately hypercholesterolemic subjects. Anthropometric parameters were also controlled.
Materials and methods
This study was conducted according to the guidelines laid down in the Declaration of Helsinki and approved by the Clinical Research Ethics Committee of the Hospital Universitario Puerta de Hierro, Majadahonda, Madrid (Spain). Experiments were performed in compliance with the relevant laws and institutional guidelines. Volunteer recruitment was carried out by placing advertisements in the Complutense University campus of Madrid and by giving short talks between lectures. The inclusion criteria for both women and men were: total cholesterol levels <200 mg/dL for the normocholesterolemic group and 200-240 mg/dL for the hypercholesterolemic subjects, non-vegetarian, non-smokers, 18 to 55 years old, not suffering from any chronic pathology or gastrointestinal disorder, and in the case of women, not pregnant. Participants should not have taken any dietary supplements, laxatives, or antibiotics six months before the start of the study and their body mass index had to be less than 30 kg/m2.
Of the fifty volunteers who enrolled in the study, 6 withdrew for personal, health or professional reasons. Of the 44 remaining volunteers, 24 were women with an average age of 25.75y (SD 6.29) and a body mass index (BMI) of 22.2 (SD 2.42) kg/m2, and 20 were men with an average age of 32y (SD 10.04) and BMI of 25.15 (SD 3.94) kg/m2. In order to avoid any bias due to gender, the number of men and women in each group was balanced; in the normocholesterolemic group, 46% were men and 54% were women, and in the hypercholesterolemic group 45% were men and 55% were women. The baseline characteristics of the 44 who completed the study, separated by group and gender, are shown in Table 1.
This was a non-randomized, controlled, crossover study of free-living individuals. After a 2 week run-in stage, subjects consumed two 200 mL servings of semi-skimmed milk per day for 4 weeks. So during this milk-intervention stage the total consumption of milk was 400 mL milk/day. Afterwards, subjects consumed two sachets of soluble cocoa powder per day in 200 mL of semi-skimmed milk, one for breakfast, and the other as a snack between lunch and dinner for 4 weeks, resulting in a total intake during this cocoa-intervention stage of 400 mL milk plus cocoa/day. The lengths of the intervention periods are in agreement with previous studies with similar objectives.14 Blood samples, blood pressure, heart rate and anthropometric measurements were taken at baseline and at the end of each intervention. The trial was held during autumn months. The commercially-available soluble cocoa product studied was provided by Nutrexpa S.L. in 7.5 g blind sachets. The cocoa dose used was established taking into account the average cocoa serving in the Spanish population,20 since the objective of the work was to reproduce the habitual consumption pattern in Spain. Furthermore, as the studied cocoa product was rich in cocoa, thus providing a strong cocoa flavour, the dose had to be one that would be largely accepted by the volunteers. For the duration of the study (i.e. the run-in and intervention stages) other cocoa products, certain fruits and vegetables (oranges, mandarins, apples, grapes, strawberries, berries in general, beets and onion), as well as their derived beverages, including wine, juices and tea, were completely restricted to reduce interindividual differences in polyphenol intake and to be able to attribute the results observed, for the most part, to the consumption of the cocoa product.
Dietary control and compliance
Apart from the aforementioned restrictions, subjects were asked to maintain the same dietary and lifestyle habits throughout the study. Dietary intake was regularly evaluated to control any possible deviations from the recommendations, and volunteers were instructed on how to fill in the dietary records before starting the study. In the last week of the run-in stage and the two intervention periods, volunteers were asked to complete a 72-hour detailed food intake report, specifying the ingredients and the amounts of food consumed, including serving weights where possible. Compliance was controlled by counting the number of cocoa servings supplied to each volunteer before and after the intervention, and by communicating with the volunteers every week. To assess dietary composition, the program DIAL (Department of Nutrition and Bromatology I. School of Pharmacy, Complutense University, Madrid) was used. The polyphenol intake was estimated using the www.phenol-explorer.eu program, and data provided by the Folin-Ciocalteu method.
Soluble cocoa powder
The commercialized soluble cocoa powder used in this study was provided by Nutrexpa S.L. together with a general analysis of its composition (6.2% fat, 12.4% proteins, 45% carbohydrates, of which 2.8% were sugar, 24.4% of dietary fibre, 4.5 % of moisture and 1.4% of minerals: 0.1% of sodium, 0.6% of calcium and 0.7% of phosphorous). A process for lowering sugar levels of cocoa products without compromising their texture had been used in the production of the cocoa product here studied.
Polyphenols in the cocoa product were extracted following a procedure designed by our group21 and analyzed spectrophotometrically as total polyphenols using the Folin-Ciocalteau reagent and gallic acid as standard. Polyphenolic composition in the extracts was characterized by high-performance liquid chromatography with diode-array detection (HPLC-DAD) using an Agilent 1200 series liquid chromatograph as described elsewhere.22 The total polyphenol concentration of the cocoa product was 34.04 ± 2.28 mg/g product and the total flavanol concentration was 3.02 ± 0.2 mg/g d.m. made up of: epicatechin 1.26 ± 0.18 mg/g d.m., catechin 0.47 ± 0.03 mg/g d.m., procyanidin B1 0.20 ± 0,04 mg/g d.m. and procyanidin B2 1.09 ± 0.10 mg/g d.m. as determined by HPLC-DAD (n=6). Therefore, the two servings of cocoa provided 43.05 mg of flavanols/day (including 18.9 mg epicatechin/day).
Total dietary fiber was 24.9% (unpublished data), and analysis was carried out following the same procedure described in Sarriá et al23.
Blood samples were drawn after 8-10 h of overnight fasting at baseline and at the end of the milk and cocoa intervention stages. Serum (without anticoagulant) and plasma (EDTA-coated tubes) were separated by centrifugation and frozen at -80ºC until analysis.
Biochemical parameter analyses
The lipid profile was determined in serum samples following reference methods or methods recommended by the Sociedad Española de Bioquímica Clínica y Patología Molecular (SEQC) using the Roche Cobas Integra 400 plus analyzer (Roche Diagnostics, Mannheim, Germany). Uric acid, creatinine and glucose were analyzed according to standardized spectrophotometric techniques and C-reactive protein (CRP) was determined using an automatized ultrasensible turbidimetric method (AU2700 Biochemistry analyser; Olympus, Watford, UK).
Inflammatory biomarker analyses
The cytokines interleukin (IL)-1, IL-6, IL-10, IL-8 and tumor necrosis factor-alpha (TNF-α) were analyzed in plasma samples using the MILLIPLEX MAP High Sensitivity Human Cytokine kit (Millipore Corp., Billerica, MA, USA). Monocyte chemoattractant protein-1 (MCP-1), vascular cell adhesion molecule-1 (VCAM-1) and intercellular cell adhesion molecule-1 (ICAM-1) were analyzed using the MILLIPEX MAP Human Cardiovascular Disease kit on Luminex equipment (Luminex-100/200, Luminex Corporation, Austin, TX, USA). High and low concentration quality controls were used with all the biomarkers. The intra- and inter-assay precision coefficients of variation were: 3.11% and 2.16%, respectively, for IL-1β; 3.51% and 4.48%, respectively, for IL-6; 3.49% and 3.78%, respectively, for TNF-α; 3.31% and 11.84%, respectively, for IL-10; 3.26% and 6.48%, respectively, for IL-8; 11.27% and 13.7%, respectively, for MCP-1; 4.5% and 8.5%, respectively, for VCAM-1; and 7.9% and 9.7%, respectively, for ICAM-1.
Antioxidant and oxidation biomarker analyses
Serum antioxidant capacity was determined by: the ferric reducing/antioxidant power (FRAP) assay,24 the free radical scavenging capacity assay using the ABTS radical cation25 and the oxygen radical absorbance capacity (ORAC) assay26 using Trolox as standard. Results were expressed as μmol of Trolox Equivalent (TE) per gram of dry matter (d.m.) of the product and as μM TE for serum samples. Serum levels of the lipid oxidation biomarker malondialdehyde (MDA) were determined as described previously,27 and carbonyl groups as biomarkers of protein oxidation were also measured.28 Serum protein was determined using the Bradford reagent.
Blood pressure measurements
Both systolic (SBP) and diastolic (DBP) blood pressure were measured using an automatic arm sphygmomanometer (Pic Indolor Diagnostic, BS 150, Artsana, Italy). At baseline and at the end of the milk and cocoa intervention stages, between 7:30 and 8:00 a.m., volunteers were asked to rest on a chair for 15 minutes before the cuff was placed on their left arm and the first reading was taken. Subsequently, two more readings were taken after 5-minute resting periods. Readings were compared and if they were not found to be within 10-15 mm Hg a fourth reading was taken.
Apart from controlling dietary intake and physical activity in order to ascertain that there were no changes in these parameters that might affect the studied cardiovascular biomarkers, anthropometric analyses were also performed as a further control. This was done in order to assess whether cocoa intervention might affect anthropometry, especially body weight and fat percentages considering their influence on cardiovascular risk. At baseline and after the two interventions, the volunteers’ total body and trunk fat percentages were assessed from tetrapolar bioimpedance measurements using a Tanita segmental body composition analyzer BC-418 MA (Tanita Corp. Tokyo, Japan). The device had a weighing system included, and height was determined using a Holtain precision mechanical stadiometer (Holtain Ltd., Crymych, Dyfed, United Kingdom). Body mass index was calculated according to the formula weight (kg)/height (m)2. Brachial, waist, abdominal, hip and thigh circumferences were measured using a SECA 203 flexible tape (SECA Ltd., United Kingdom). Tricipital and subscapular skin folds were measured using a Harpenden skin fold caliper (Holtain Limited Company, Crosswell, Crymych, United Kingdom). By means of these biometric data, body density29 and the percentage of body fat30 were calculated.
Physical activity analysis
Participants were asked to maintain their usual level of physical activity during the study. Volunteers filled out a questionnaire before starting the study in order to know their occupation and leisure time activities, and consequently the physical activity involved. In the run-in and the two intervention stages, physical activity was calculated using an adapted version of the Minnesota Leisure Time Physical Activity Questionnaire.31 We assumed that 1 metabolic equivalent is approximately 1 kcal/min for a 70-kg man. As all the volunteers’ occupations involved low physical activity, their energy expenditure was calculated by only taking into account leisure time. Data was expressed as kcal/day.
Taking total cholesterol as the main variable, an average size of 23 subjects per group was calculated in order to obtain a statistical power of 80 percent that the study would detect a treatment difference at 0.05 significance level, if the true difference between treatments was 6 mg/dL. This was based on the assumption that the within-patient standard deviation of the response variable would be 10.
Data have been presented as means ± standard error of the mean, unless specified otherwise. Before any statistical analysis, all variables were examined for normality using the Kolmogorov-Smirnov test. A two-factor repeated measures factorial design was used to examine the effects due to time (treatment), and group and time*group interaction. Since the effect on the group and the dietary treatment (time)*group interaction were not significant in any of the parameters analyzed, these statistical results are not included in tables.
In order to avoid any bias due to gender, the number of men and women in each group was balanced: 46% were men in the normocholesterolemic group, and 45% in the hypercholesterolemic group. Differences within either the normocholesterolemic or hypercholesterolemic groups were studied using the Bonferroni post-hoc test. Statistical significance was set at p<0.05. The SPSS statistical package (version 19.0; SPSS, Inc., IBM Company) and Statgraphics Centurion XVII (Stat Point Technologies, Inc.) were used.
The 72-hour intake reports provided by volunteers were pre-filtered to exclude any considered unreliable, which were defined as those that showed daily energy intake estimations 70% below light activity energy intake recommendations or 130% above normal activity energy intake recommendations, taking into consideration the age and gender of the group.20 The analysis of the food reports showed that carbohydrate (p=0.002) and protein intake (p<0.001) were statistically higher after consuming the cocoa product with milk (Table 2). According to the paired test, in the normocholesterolemic group the protein intake was significantly higher after the cocoa intervention stage compared to baseline. Also, dietary fiber increased after this stage (p=0.050) due to the dietary fiber provided by the cocoa product, even though recommended intakes were not reached.32 The intake of polyphenols also increased in both groups after the cocoa intervention stage because of the phenolic content of the soluble cocoa product, though only slightly (not statistically significant) owing to the restriction on polyphenol-rich foods. None of the other dietary intake parameters showed statistical differences due to the dietary treatment.
After consuming the cocoa product, total cholesterol, LDL-C and triacylglycerol levels did not show any statistical differences, whereas HDL-C had increased significantly (p<0.001; Table 3). In both groups, the HDL-C values achieved after the interventions were higher than their respective baseline concentrations according to the paired tests. Both alanine and aspartate aminotransferase levels increased with respect to baseline in normo- and hypercholesterolemic groups after the intake of cocoa and milk, according to the repeated measures factorial design (p=0.001) and the Bonferroni test. However, both parameters remained within their respective reference ranges of normality established by SEQC (0-41 and 0-38 U/L, respectively). In contrast to glucose and uric acid levels, which did not change, urea (p=0.009) increased in both groups when compared to baseline after the milk and cocoa interventions, and creatinine (p=0.047) was higher in both groups after the milk intervention with respect to baseline and the cocoa intervention. Nevertheless, both urea and creatinine values remained within their reference ranges according to SEQC (10–50 mg/dL and 0.50-1.30 mg/dL, respectively).
Levels of inflammatory and adhesion molecules varied considerably (Table 4). IL-1β, IL-6 and IL-8 were within the levels described for healthy subjects (the control group),33 whereas IL-10 values were above the corresponding levels (2.4-6.6 pg/mL) and TNF-α values below (14.2-61.7 pg/mL). IL-1β, TNF-α and MCP-1 concentrations decreased compared to baseline values after consuming the cocoa test beverage, particularly in the hypercholesterolemic group, but without reaching the level of statistical significance. On the other hand, IL-10 levels were significantly lower (p=0.022) after the cocoa intervention.
VCAM-1 values in the normocholesterolemic group were close to the higher limit of the range (46–166 ng/mL) described for control subjects,33 and in the hypercholesterolemic group values were above the upper limit at baseline. In contrast, ICAM-1 values in both groups were within the range of control subjects (39–79 ng/mL),33 being higher in the hypercholesterolemic group compared to the normocholesterolemic group. After the intervention stages, there was a decrease of VCAM-1 in both groups, without reaching the level of statistical significance (Table 4).
At the start of this study, subjects were normotensive (SBP<140 mmHg and DBP≥80 mmHg) with the hypercholesterolemic group registering higher blood pressure values, though not significantly different from those of the normocholesterolemic group. The interventions did not induce any changes in these blood pressure values or heart rate (Table 5).
Antioxidant capacity (as measured by the ORAC, FRAP and ABTS methods) and levels of protein (carbonyl groups) and lipid (MDA) oxidation showed no significant differences after the interventions (Table 5). There were no changes either in any of the anthropometric parameters (Table 6) or in body density or percentage of body fat (data not shown). Likewise, there were no changes in energy expenditure values, which were 925 ± 101, 835 ± 153 and 848 ± 100 kcal/d at baseline and for milk and cocoa interventions, respectively, in the normocholesterolemic group, and 622 ± 245, 534 ± 230 and 569 ± 168 kcal/d at baseline and in milk and cocoa interventions, respectively, in the hypercholesterolemic group.