Chocolate and Cardiovascular Health
By Craig Schneider, MD, and Elizabeth Strawbridge, MD
Dr. Schneider is Director of Integrative Medicine, Department of Family Medicine, Maine Medical Center in Portland. Dr. Strawbridge is a graduate of the Integrative Family Medicine Fellowship at Maine Medical Center and the Fellowship at the Arizona Center for Integrative Medicine. She is in private practice.
Dr. Schneider and Dr. Strawbridge report no financial relationships relevant to this field of study.
Observational studies suggest that dietary flavonoids (catechins, epicatechins, and oligomeric proanthocyanadins) reduce the risk of death from cardiovascular disease.1 Among the often delicious flavonoid options of red wine, teas, fruits, and vegetables, chocolate has a special place in the hearts of many.
Native to South America, chocolate is produced from the seed of the small tropical evergreen tree, Theobroma cacao. Cultivated by Mesoamerican peoples for thousands of years, the bitter seeds are removed from ripe pods, fermented, dried, cleaned, and then roasted. After roasting, the shell is removed from the seeds to produce cacao nibs, which then may be ground into cocoa mass. Cocoa mass, which is rich in bitter flavanols, is liquified and divided into cocoa solids and cocoa butter. These are then turned into a variety of culinary chocolates via alteration of their ratios and the addition of various sweeteners, fats, milk, and other ingredients. White chocolate contains no cocoa solids but only cocoa butter, milk, and sugar. Natural cocoa powder (undutched) is light brown in color. Dutch-processed cocoa has been alkalinized, which reduces bitter flavors and increases smoothness, resulting in a darker color at the expense of reducing the amount of active flavanols by up to 60%.2 Thus, processing techniques influence flavanol content significantly; a product’s cocoa content and color are not necessarily accurate reflections of flavanol content.3 A very dark colored chocolate may be low in flavanols.3
Mechanism of Action
Cocoa contains oils, tannins, and alkaloids such as theobromine and caffeine, but the potential cardiovascular benefits are believed to be primarily related to its major flavonoids from the flavanol subgroup, epicatechin and catechin. In vivo and ex vivo studies demonstrate chemically pure epicatechin closely emulates the acute vascular effects of flavanol-rich cocoa by improving endothelial function and increasing production of nitrous oxide.4 Sustained improvements in endothelial dysfunction by regular dietary intake of flavanols from a cocoa-rich beverage also are associated with mobilization of functional circulating angiogenic cells, also known as endothelial progenitor cells, which are critical for vascular repair and maintenance of endothelial function.5 In multiple trials, cocoa and chocolate both reduced insulin resistance as measured by Homeostasis Model of Assessment-Insulin Resistance (HOMA-IR) and improved endothelial function as measured by flow mediated dilation (FMD) through improved formation of vasodilating nitric oxide.6
A meta-analysis of randomized controlled trials published in The American Journal of Clinical Nutrition in March 2012 looked exclusively at the effects of chocolate, cocoa, and flavan-3-ols on cardiovascular health.6 The meta-analysis included 42 “acute” or “short-term chronic” (< 18 week) randomized, controlled trials (RCTs) comprising almost 1300 patients randomly assigned to chocolate, cocoa, or cocoa flavan-3-ol preparations in comparison with control groups. Major cardiovascular disease (CVD) risk factor outcome findings included: 1) significant reductions in serum insulin levels suggesting improvement in insulin resistance HOMA-IR (-0.67; 95% confidence interval [CI], -0.98 to -0.36); 2) improvement in endothelial function based on FMD after both chronic (1.34%; 95% CI, 1.00% to 1.68%) and acute (3.19%; 95% CI, 2.04% to 4.33%) ingestion; 3) reduction in diastolic blood pressure (DBP) (-1.60 mmHg; 95% CI, -2.77 to 0.43 mmHg) and mean arterial pressure (-1.64 mmHG; 95% CI, -3.27 to -0.01 mmHg); and 4) marginally significant effects on LDL (-0.07 mmol/L; 95% CI, -0.13 to 0.00 mmol/L) and HDL cholesterol (0.03 mmol/L; 95% CI, 0.00 to 0.06 mmol/L). Although it appears FMD improved regardless of cocoa dose, it appears only higher doses (> 50 mg epicatechin/day) reduced systolic blood pressure (SBP) and DBP. Two-thirds of a 100 g Ritter Sport Halbitter Bar contains 44 mg of epicatechin. Although the health outcomes were positive overall, it is important to note that the GRADE assessment of strength of evidence varied from low to moderate. The quality of these studies suffers due to small sample size, unclear reporting of allocation concealment and dropouts, as well as missing outcomes data. A majority of these studies were industry-funded, thus introducing strong risk of bias. However, even after sensitivity analysis removed industry-funded studies, the results were not altered with regard to cocoa’s effects on insulin resistance and FMD, although other outcomes lost statistical significance.
Another recent meta-analysis supports the positive findings of chocolate’s benefits on blood pressure (BP).7 This meta-analysis included 20 short-term (mostly 2-week) studies involving 856 healthy participants demonstrating an SBP decline of -2.8 mmHg (95% CI, -4.72 to 0.82; P = 0.005) and DBP decline of -2.2 mmHg (95% CI, -3.46 to 0.93; P = 0.006). This effect was more significant in the shorter trials as compared to longer trials, although reasons for that are unclear and may be confounded by other factors. Subgroup analysis revealed that a significant BP lowering effect was found only in trials where the control groups had flavanol-free product vs low-flavanol product; 78% of the short (< 2 week duration) studies used a flavanol-free product. Additionally, the dose of chocolate was higher in the shorter studies (100 g/d). However, when looking at total flavanol content in the active group, there was no correlation between dose of flavanol (between 3.6-105 g/day) or specifically epicatechin and catechin (between 7-236 mg/d) and BP. Nor was there an association between theobromine content and BP. In conclusion, BP improved slightly with short-term ingestion of varying amounts of chocolate and more markedly when compared with control groups that were flavanol-free (i.e., white chocolate) vs low flavanol (i.e., heavily dutched or milk chocolate). Once again, it appears longer studies investigating chronic, low-dose cocoa ingestion are needed to better understand cocoa’s utility in BP control.
The following individual studies that were presented in both of the above meta-analyses may further help inform our view of cocoa and chocolate products.
In 2005, Grassi et al performed a randomized, controlled, crossover study of 20 untreated hypertensive adults and 15 normotensive controls randomly assigned to eat one full 100 gram, commercially available, Ritter Sport Halbbiter dark chocolate bar containing 88 mg flavanols (21.91 mg catechin, 65.97 mg epicatechin) or a flavanol-free isocaloric Kraft Milka white chocolate bar per day for 15 days, followed by a 7-day washout, and followed by crossover to the other treatment.8 The participants were instructed to avoid other flavanol-rich foods throughout the study. Baseline FMD in hypertensives was impaired compared to controls 7.4 vs 9.9, but improved in hypertensives to almost normal following the dark chocolate intervention (8.9 ± 1.4%; F = 13.25; P < 0.0001) but NOT the flavanol-free white chocolate (7.5 ± 1.3%; P < 0.0001). Normal FMD can vary from lab to lab. The studies do not specifically mention CI, but do include F values, which compare variance (ANOVA). FMD also increased in controls after dark chocolate (11.8 ± 1.3%; F = 19.86; P < 0.0001) but not white chocolate (10.1 ± 0.9%, NS). Additionally, ambulatory BP decreased in the dark chocolate hypertensive group (24-hour SBP -11.9 ± 7.7 mm Hg; F = 33.78; P < 0.0001; 24-hour DBP -8.5 ± 5.0 mmHg; F = 38.80; P < 0.0001), but not in the white chocolate group. Dark but not white chocolate also decreased fasting insulin levels (from 13.1 to 9.3 mUI/mL; F = 33.55; P < 0.0001).
In 2007, Taubert et al performed a similar RCT with 44 pre-hypertensive or untreated stage I hypertensive patients.9 This study used a lower-dose habitual cocoa intake over a longer period of time (18 weeks vs 15 days in the Grassi study). Again the commercially available Ritter Sport Halbitter was used as the intervention while a matching isocaloric Kraft Milka served as control. Instead of the 100 g bar, only 6.3 g was used. Despite this cocoa dose being one-sixteenth that used in Grassi’s study, translating to only 5.1 mg epicatechin/day, there were still improvements in BP. From baseline to 18 weeks, the dark chocolate group’s SBP decreased by -2.9 mmHG (95% CI; P < 0.001) and DBP by -1.9 (95% CI; P < 0.001) without changes in body weight or glucose. White chocolate intake caused no changes in BP. This study did not assess insulin sensitivity or endothelial function.
There has been concern that even if benefits are proven for cocoa, daily ingestion of the sugar and fat associated with commercially available cocoa products might adversely affect cardiovascular health through caloric impact and role in inflammation. It is thus surprising that blood sugar actually was lowered following sugar-containing cocoa product ingestion in a number of studies. In the Grassi study, participants were instructed to substitute the 480 kcal chocolate bars for foods of similar energy composition in an attempt to avoid weight gain,8 but the Njike et al study comparing sugar-sweetened vs sugar-free cocoa on endothelial function in overweight individuals neither demonstrated weight gain nor significant differences in benefit between groups over a 6-week period.10 Participants were instructed to maintain their usual diet without altering their intake to adjust for the additional chocolate calories. They were, however, told to avoid flavonoid-rich foods 24 hours prior to test days.
Forty-four adults were included in this randomized, controlled, crossover trial where they were sequenced into a twice daily sugar-free cocoa beverage (containing a total of 22 g/d cocoa, 69 mg of combined catechin and epicatechin, 90 kcal), a sugar-sweetened cocoa beverage (22 g/d cocoa and 91 g/d sugar, 460 kcal containing same amount flavanols as the sugar free), or a sugar-sweetened, cocoa-free placebo (110 g/d sugar without cocoa, 500 kcal, no flavanols). Each treatment was administered daily for 6 weeks, with a 4-week washout period before groups were switched to the next intervention. The results revealed that consumption of cocoa-containing beverages for 6 weeks improved FMD compared to cocoa-free placebo. The magnitude of improvement in FMD after consumption of the sugar-free preparation was greater than after sugar-sweetened cocoa but not significantly (P = 0.15). Overall, FMD was 3.2% greater in sugar-free cocoa vs placebo (95% CI, -1.8 to 4.6, P < 0.001) and 2.3% greater in sugar-sweetened cocoa vs placebo (95% CI, 0.9 to 3.7, P = 0.002). Other biomarkers of cardiac risk (BP, lipids, fasting glucose, C-reactive protein) did not change appreciably from baseline.
Although many of these studies demonstrate cocoa’s beneficial effects on CVD risk biomarkers (insulin levels, FMD, nitric oxide levels, BP), they do not attempt to elucidate chocolate’s role in actual cardiovascular mortality. What about the impact of cocoa in populations? A retrospective epidemiological study of the Kuna, a group of people living off the coast of Panama known to have substantially lower BPs as they age compared to their mainland relatives in Panama, provides an opportunity to examine cocoa intake and mortality.11 The Kuna drink a flavanol-rich cocoa as their main beverage, contributing to an intake of 900 mg flavanol per day (likely the most flavanol-rich diet of any population). Bayard et al reviewed causes of death from 2000 to 2004 on mainland Panama vs San Blas islands where the Kuna live. On San Blas, death from ischemic heart disease had a frequency of approximately 8 per 100,000/year and stroke of 3 per 100,000/year, compared to mainland Panama, where ischemic heart disease was 44 per 100,000/year and stroke was 41 per 100,000/year. Because this is an observational study and many confounders are surely at play, firm conclusions cannot be drawn; however, it is suggestive of the beneficial impact of cocoa on a population and supports the findings of many (but not all) small, randomized, clinical trials looking at CVD risk factors.
The Zutphen Elderly study also sought to examine cocoa intake on a real population. This prospective, cross-sectional, cohort study assessed habitual cocoa intake and BP in 470 elderly Dutch men over 5 years.12 The lowest tertile comprised no cocoa intake. The median intake in the middle tertile was 0.92 g/d cocoa and in the highest tertile was 4.18 g/d. Plain chocolate and chocolate bars contributed two-thirds of cocoa intake; the rest came from chocolate candies, spreads, mousse, pudding, and hot cocoa drinks. Adjusted for confounding factors (age, body mass index, socioeconomic status, cigarette smoking, alcohol, physical activity, use of aspirin and anticoagulants, drug-use, food [vegetables, fruit, low and medium fat dairy, meat, non-chocolate sugary confections, nuts, seeds, coffee], and total calorie intake), there was lower mean SBP (-3.7 mmHG; 95% CI, -7.1 to -0.3) and DBP (-2.1 mmHg; 95% CI, -4.0 to -0.2) in the highest tertile of cocoa intake compared with the lowest. After 15 years of follow-up, 314 men had died, 152 of cardiovascular diseases. Compared with the lowest tertile of cocoa intake, the adjusted relative risk for men in the highest tertile was 0.50 (95% CI, 0.32 to 0.78; P = 0.004 for trend) for cardiovascular mortality and 0.53 (95% CI, 0.39 to 0.72; P < 0.001) for all-cause mortality. This translates to an impressive risk reduction of about 50% for both cardiovascular and all-cause mortality associated with regular, low-dose, commercial cocoa-containing products in this mortality based study on real-life cocoa consumption in a Western population.
Cocoa is well-tolerated (nay, extremely well-tolerated) orally. There are reports of allergic skin reactions, shakiness, increased urination, increased heart rate, stomach upset, gastroesophageal reflux, nausea, and constipation, as well as triggering of migraine headaches in certain individuals.13 Weight gain may be a concern because of the high sugar and fat content of some cocoa and chocolate products but as discussed above studies cast doubt on this assumption.10
Although popular media has flaunted chocolate’s health benefits, and we would love to be able to recommend unlimited chocolate for our patients, the randomized trials that have been done so far do not conclusively prove that chocolate will reduce one’s risk of CVD. They do, however, demonstrate acute and chronic benefits of cocoa and chocolate on FMD and insulin resistance, and possibly but less conclusively improvements in BP (BP changes were documented, but not really clinically relevant). Although there is much variability in the study designs and populations, there seems to be clear uniformity in endothelial function improvement regardless of study length, cocoa strength, or population. None of the studies reviewed demonstrate a negative impact of cocoa or chocolate intake; however, firm conclusions remain elusive regarding their improvement on BP and fasting glucose due to mixed findings and study limitations.
Because the impact of even small improvements in these areas from a dietary intervention that is both well-tolerated and adhered to could translate into substantial improvements in morbidity and mortality at a population level, larger and longer-term independently funded trials are needed for confirmation of their impact on outcomes that matter clinically. Luckily, due to gourmand health enthusiasts worldwide, more and more studies are emerging that explore whether this guilty pleasure could instead be a healthful indulgence.
One challenge is finding a product that reliably delivers adequate doses of flavanols without significantly increasing caloric burden and weight gain, which might limit the potential cardiovascular benefits (although this is called into question by the Njike study, which found no weight gain in the group consuming sugar-sweetened, higher-calorie chocolate compared to sugar-free, calorie-reduced chocolate and benefits were similar.)10
Many chocolate studies have used proprietary flavanol-enhanced cocoa drinks or chocolate, which makes it difficult to determine practical recommendations for patients. There is concern that many commercially available cocoa products are not rich in flavanols. In addition, there are few means for consumers to be able to gauge flavanol content reliably.
Because several studies used Ritter Sport Halbitter chocolate (50% cacao bittersweet) 100 g (~480 calories, 88 mg of cocoa flavanols), this might be a reasonable product to recommend. In studies, both 100 grams and 6.3 grams performed well. Until conclusive studies demonstrate no risk of weight gain associated with chocolate candy intake, it seems prudent to discuss strategies for substituting this for foods of similar energy composition in order to not gain weight and potentially counteract the potential benefits obtained by chocolate consumption. It also may be useful to remind patients that cocoa powder is defatted and thus quite a bit lower in caloric content than chocolate. Unfortunately, the flavanol content of commercially available cocoa drinks are not known and products used in studies do not appear to be commercially available. Mars makes CocoaVia, which guarantees 250 mg of cocoa flavanols from cocoa extract in several flavors to be mixed into drinks (low calorie 20-30 Kcal per serving), but we were unable to find clinical trials utilizing this preparation. They are currently recruiting for a study using this product examining the role of flavanols in Cardiovascular Function in Healthy Aging at the University of Texas in Austin.
In summary, enjoy chocolate — not too much, mostly dark, never white — and bask in the knowledge that it is probably good for you.
1. Hertog MG, et al. Flavonoid intake and long-term risk of coronary heart disease and cancer in the seven countries study. [Erratum appears in Arch Intern Med 1995;155:1184]. Arch Intern Med 1995;155:381-386.
2. Andres-Lacueva C, et al. Flavanol and flavonol contents of cocoa powder products: Influence of the manufacturing process. J Agricul Food Chem 2008;56:3111-3117.
3. Hollenberg NK, Fisher ND. Is it the dark in dark chocolate? Circulation 2007;116:2360-2362.
4. Schroeter H, et al. (-)-Epicatechin mediates beneficial effects of flavanol-rich cocoa on vascular function in humans. Proc Natl Acad Sci U S A 2006;103:1024-1029.
5. Heiss C, et al. Improvement of endothelial function with dietary flavanols is associated with mobilization of circulating angiogenic cells in patients with coronary artery disease. J Am Coll Cardiol 2010;56:218-224.
6. Hooper L, et al. Effects of chocolate, cocoa, and flavan-3-ols on cardiovascular health: A systematic review and meta-analysis of randomized trials. [Review]. Am J Clin Nutr 2012;95:740-751.
7. Ried K, et al. Effect of cocoa on blood pressure. Cochr Database System Rev 2012;8.
8. Grassi D, et al. Cocoa reduces blood pressure and insulin resistance and improves endothelium-dependent vasodilation in hypertensives. Hypertension 2005;46:398-405.
9. Taubert D, et al. Effects of low habitual cocoa intake on blood pressure and bioactive nitric oxide: A randomized controlled trial. JAMA 2007;298:49-60.
10. Njike VY, et al. Effects of sugar-sweetened and sugar-free cocoa on endothelial function in overweight adults. Int J Cardiol 2011;149: 83-88.
11. Bayard V, et al. Does flavanol intake influence mortality from nitric oxide-dependent processes? Ischemic heart disease, stroke, diabetes mellitus, and cancer in Panama. Int J Med Sci 2007;4:53-58.
12. Buijsse B, et al. Cocoa intake, blood pressure, and cardiovascular mortality: The Zutphen Elderly Study. Arch Intern Med 2006; 166:411-417.
13. Vlachopoulos C, et al. Effect of dark chocolate on arterial function in healthy individuals. Am J Hypertens 2005;18:785-791.