Share this article on:

Vascular Action of Cocoa Flavanols in Humans: The Roots of the Story

K. Hollenberg, Norman MD, PhD

Journal of Cardiovascular Pharmacology: June 2006 - Volume 47 - Issue - p S99-S102
Original Articles

Diet patterns are widely recognized as contributors to hypertension. Widely studied potential contributors include intake of sodium, potassium, magnesium, calcium, soluble fiber, ω-3 fatty acids, alcohol, protein, and calories. We add to that list the effect of dietary flavanols present in certain cocoas, which have sufficient activity on vascular nitric oxide to influence blood pressure control. Kuna Indians who live on islands near Panama have little age-related rise in blood pressure or hypertension. On migration to Panama City, blood pressure rises with age, and the frequency of essential hypertension matches urban levels elsewhere. We have identified a specific food that probably makes an important contribution to cardiovascular status. Island-dwelling Kuna drink more than 5 cups of flavanol-rich cocoa per day and incorporate that cocoa into many recipes. Mainland Kuna ingest little cocoa, and what they take is commercially available and flavanol-poor. The flavanol-rich cocoa activates nitric oxide synthase in vitro and in intact humans in the doses that the Kuna employ. Vasodilator responses to flavonoid-rich cocoa are prevented or reversed by the arginine analog, N-nitro-L-arginine methyl ester. Island-dwelling Kuna have a 3-fold larger urinary nitrate:nitrite than do Mainland dwellers. As endothelial dysfunction is central to current thinking on cardiovascular pathophysiology, a food that enhances endothelial function could have broad implications. The list of candidate conditions that might be influenced is impressive, ranging from atherosclerosis and diabetes mellitus to hypertension and preeclampsia, to vascular dementias and end-stage renal disease. The next decade will be interesting.

Department of Medicine and Radiology, Brigham and Women's Hospital and Harvard Medical School, Boston, MA

Reprints: Norman K. Hollenberg, MD, PhD, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115 (e-mail:

The Kuna Indians live on offshore islands on the Caribbean Coast of Panama and have probably lived there for centuries, since the time of the Spanish conquistadors over 500 years ago.1 The islands are rocky, wind swept, and dry. Hence, there are no mosquitoes. As their major health problems on the mainland involve malaria, yellow fever, and dengue, and as each has the mosquito as the major vector, these islands must have seemed like paradise. Moreover, a barrier reef separates most of the San Blas Archipeligo from the ocean and, thus, the islands were inaccessible to Spanish Galleons. The Kuna had found a safe heaven. About 60 years ago, Kean wrote an article in which he described his findings on blood pressure in the Kuna. He noted that hypertension was very rare, and blood pressure did not rise with age.2

This project began more than a dozen years ago with an original goal of identifying genes that protect against high blood pressure. One strategy for identifying such genes involves identifying a geographically isolated, ethnically homogeneous group of individuals—ideally the product of a small founder colony who have lived in isolation for generation, and who have inbred.3

I found the Kean article as the product of a specific and focused search. Gordon Williams and I had been studying a subset of patients with essential hypertension that we call “nonmodulators.”4 Compelling evidence for a contribution of genetics included a striking family history of hypertension in nonmodulators and concordance in sibling pairs of some of the features.4 We have identified genes that predispose to nonmodulation, specifically polymorphisms involving the angiotensinogen gene locus.5,6 We have also gradually come to understand what the term “polygenic” means. What it means is that any gene polymorphism will account for only a small fraction of the hypertension. Far more satisfying is to identify a major gene effect and study that.

A logical possibility that had not been systematically assessed involves protective genes. If there are “bad genes” that can promote hypertension, might there not be “good genes” that protect against hypertension? In about 1990, I decided to make that a focus of my research efforts.

The search strategy that was developed was straightforward. If one could identify geographically isolated groups of people, ideally having lived in isolation for many generations; and if they had taken protective genes into their geographic isolation with them at the time their colony was created; and if they inbred during many generations, one might anticipate finding a population that had a number of important features: Hypertension would be uncommon and blood pressure would not rise with age. This was precisely what Kean had described in Kuna decades earlier.

When we obtained blood pressure measurements in island-dwelling Kuna, we were very enthusiastic. Hypertension was extremely uncommon, and blood pressure did not rise with age (Fig. 1). This did not reflect a low-salt diet. Indeed, the intake of sodium and chloride was rather larger than in most Western populations.7



The next step in the story demanded that we test the hypothesis that they carry protective genes. If they were, indeed, genetically protected, one would anticipate that they could move to an urban environment and still remain normotensive. We identified several hundred Kuna living either in Panama City proper or in an adjacent suburb. To our chagrin, the immigrants to an urban environment enjoyed all of the benefits of modern, Western urban life (Fig. 2). Hypertension became substantially more common, and blood pressure rose significantly with age. They were not protected by genes. Whatever was keeping their blood pressure down was environmental. It was also clear that it was not sodium intake.7



Next, we undertook a systematic assessment of renal perfusion and function in indigenous island-dwelling Kuna.8 The rationale for this decision had several elements. First, renal perfusion and function are intrinsically interesting in relation to hypertension, or its absence. Perhaps more fundamental was the fact that research funding was needed urgently if this project was to continue. The Baxter Foundation agreed to support a study designed to test a specific relevant hypothesis. One thought at that time held that at least part of normal human renal aging reflected the combined effects of the blood pressure rise with age that occurs with apparent normal aging in humans and a high protein intake. The Kuna living in their indigenous island setting did not have a blood pressure rise, and protein intake was rather low. Thus, we could test that hypothesis in humans.

As has been shown many times before, we identified a significant downward slope in the relation between both polycyclic aromatic hydrocarbon (PAH) and inulin clearance and age in volunteers in Boston. Using the same PAH and inulin source and pumps essentially identical to those used in Boston, we examined renal perfusion and function in the Kuna.8 Both PAH clearance and inulin clearance fell with age in the Kuna, and the slope was steeper than that in Boston. Their kidneys were not protected from aging. Of specific interest was the finding that the kidneys were widely vasodilated in the Kuna. Renal plasma flow and especially glomerular filtration rate were substantially higher in the Kuna than in age-matched whites in Boston. At that time we mused on a possible role for nitric oxide, which is one of the few mediators in the body that will produce these renal hemodynamic findings.9

Much of the budget provided by the Baxter Foundation went into a systematic assessment of the diet of island-dwelling Kuna.10 A number of differences that were potentially relevant were found. Urinary sodium, magnesium, and calcium did not differ between island dwellers and immigrants to the suburbs. Thus, it is unlikely that intake differed. Urinary potassium in the island setting (48±3 mEq/g creatinine) was significantly greater than that in the mainland (41±2 mEq/g creatinine; P<0.05). This difference, however, is well below the amount of potassium usually given as a supplement in potassium-loading studies designed to assess the influence of potassium on blood pressure. Indeed, a dose 7 times higher has a modest effect on blood pressure, certainly not returning blood pressure to 110/70 mm Hg and abolishing hypertension completely—as seen in the island dwellers. The difference in potassium excretion presumably reflected the fact that island dwellers ate substantially more fresh fruit including plantains, green bananas, and mangos.

As reviewed in detail by McCullough et al later in this symposium, protein intake was modest in all locations. Fish ingestion, however, was substantially greater in the island community than in the mainland. There have been suggestions that ω-3 fatty acids in fish can have an important cardiovascular influence, including an effect on blood pressure, but once again, the claims on blood pressure involve a modest effect. Blood pressure is not returned to low-normal levels in everyone with fish ingestion.

The most outstanding finding was the fact that most of the island-dwelling Kuna drank cocoa as their major drink and did so every day. The quantities of cocoa ingested were very substantial. Indeed, for many it was their only source of drinking water. Our estimate of cocoa intake, averaging 35 cups per week per individual Kuna, is almost certainly an underestimate. For many Kuna, the only drink is cocoa, and in that hot and humid climate they probably drink more than 5 cups per day.

At that juncture, we began a collaboration with Mars, Incorporated who provided funding for the project, but—at least as important—relevant laboratory support in the form of food composition analysis. They quickly demonstrated a crucial finding. The cocoa ingested by the Kuna is naturally very rich in a specific subclass of flavonoids known as flavanols, and especially (−)epicatechin, (+)catechin, and flavanol-based oligomers known as procyanidins.10,11 We also learned from this collaboration that commercially available cocoa-based products, what we can buy in stores, is routinely very low and/or inconsistant in flavanol content compared with what was observed with the Kuna cocoa as a result of routine postharvest handling and food processing techniques that are implemented to improve taste and appearance of cocoa-based products in conformance with the preferences of many consumers in Western Europe and the United States.11 In addition, we were made aware of, and subsequently a part of, a research collaboration between Mars and the University of California, Davis. That team, led by Carl L. Keen, had used various cell and tissue culture systems to study purified and well-characterized flavanol fractions from cocoa isolated by the Mars chemists. Key observations made in this research included a striking ability of certain cocoa flavanol fractions to induce synthesis of prostacyclin and, perhaps even more exciting, relaxation of isolated aortic rings in a manner consistent with increased nitric oxide synthesis.16,17

Spurred by these observations, we undertook a study of the vascular actions of a specific well-characterized flavanol-rich cocoa in Boston.12 As detailed elsewhere in this proceeding, we found a striking flavanol-induced activation of vascular nitric oxide synthesis in healthy humans. As an essentially identical flavanol-poor cocoa induced a much smaller response, clearly the flavanols were largely responsible.12

The mechanistic sequence remains obscure, but we have made observations that may be relevant. The response to the first dose of flavanol-rich cocoa is not at maximum. The response becomes larger until it reaches a new plateau: From our preliminary data, it would seem that 5 to 7 days are required to reach a new steady state. Moreover, and possibly related, a residual vascular response is evident 15 hours after the last dose, at a time when pharmacokinetic studies indicate that the flavanols and their known metabolites have largely disappeared from the circulation.11 Thus, one possible mechanistic sequence would involve activation at the nitric oxide synthase gene level, as a first step. As a second possibility the responsible agent may be a metabolite of the flavanols that accumulates gradually.

Our recent observations indicate that flavanol-rich cocoa influences vascular function not only in the extremity and in the kidney but also in the blood supply to the brain presented elsewhere in this symposium. Vascular function was improved not only in the healthy young human, but also in the healthy human that had reached 70 or 80 years of age. Thus, the endothelial dysfunction that occurs with normal aging seems to be functional rather than organic in nature. Equally exciting is the observation that the patient with type 2 diabetes mellitus and endothelial dysfunction shows reversal of endothelial dysfunction within a week of beginning the daily ingestion of flavanol-rich cocoa. If this observation made in the extremity extends to the heart, brain, and kidneys, the excitement will be substantial. All of these observations have clinical implications.

Our long-term plans involve extensions of these physiological observations in natural history studies. If the brain blood flow study continues to show an affect, it will be interesting to see whether flavanol-rich cocoa will have a positive influence on the development of vascular dementias, as discussed later in the symposium. If reversal of endothelial dysfunction in the patient with diabetes mellitus proves to be a frequent event, it would be interesting to see whether flavanol-rich cocoa will exert a positive influence on atherosclerotic disease and diabetes mellitus or the development of end-stage renal disease. We have already shown in preliminary studies a positive influence of flavanol-rich cocoa on renal perfusion in the patient with diabetes mellitus.

An especially attractive target, we believe, involves the young woman with preeclampsia. Because of a socioeconomic background, one anticipates in young Kuna primipara a frequency of preeclampsia that exceeds 10% and may approach 20%. In fact, preeclampsia is very uncommon in the Kuna. In Africa, a young woman is far more likely to develop preeclampsia than the 5% figure usually cited in the western world and is 200 times more likely to die of her complicated pregnancy. Multiple observations suggest that the pathogenetic sequence in preeclampsia begins with an inadequate placenta.13 As a consequence, there is a systemic effect that leads to a rise in blood pressure, proteinuria, and other manifestations of this high-risk state. Multiple observations suggest that nitric oxide deficiency underlies this systemic vascular effect. Flow-mediated dilatation, widely demonstrated to reflect nitric oxide, is increased in the normal pregnancy but reduced in the woman with preeclampsia.14 That reduction in flow-mediated dilatation shows remarkable correlation with the accumulation in plasma of asymmetric dimethyl-arginine, an endogenous inhibitor of nitric oxide.15 There is currently no effective preventative treatment for preeclampsia. If the cocoa flavanols were a drug, we would need a great deal of experience before proposing a clinical trial, especially in the pregnant woman. On the other hand, cocoa has been part of the human culture for over 2000 years and is unlikely to be harmful.

We look forward to the next symposium in the hope that by that time we will be able to describe the study design and the initiation of one or more natural history studies.

Back to Top | Article Outline


The work was supported by the Mars, Incorporated, National Institutes of Health Grants T32 HL-076909, NCRR GCRC M01RR026376, and 1 P50 ML53000-01.

Back to Top | Article Outline


1. Howe J. The Kuna Gathering. Contemporary Village Politics in Panama. Austin, TX: University of Texas Press; 1986.
2. Kean BH. The BP of the Kuna Indians. Am J Trop Med Hyg. 1944;24:341–343.
3. James GD, Baker PT. Human population biology and blood pressure: evolutional and ecological considerations and interpretations of population studies. In: Laragh JH, Brenner BM, eds. Hypertension, Pathophysiology, Diagnosis, and Management. New York: Raven Press; 1995:115–125.
4. Hollenberg NK, Williams GH. Abnormal renal function, sodium-volume homeostasis and renin system hypertension: the evolution of the nonmodulator concept. In: Laragh JH, Brenner BM, eds. Hypertension: Pathophysiology, Diagnosis, and Management. New York: Raven Press; 1995:1837–1856.
5. Koshachuhanan N, Hunt SC, Hopkins PN, et al. Genetic determinants of nonmodulating hypertension. Hypertension. 2003;42:901–908.
6. Williams GH, Hollenberg NK, Hopkins PM, et al. The role of intermediate phenotypes in essential hypertension: non-modulation as a model. Endocrine Res. 1996;22:675–680.
7. Hollenberg NK, Martinez G, McCullough M, et al. Aging, acculturation, salt intake, and hypertension. Hypertension. 1997;29:171–176.
8. Hollenberg NK, Rivera A, Meinking T, et al. Age, renal perfusion, and function in island-dwelling indigenous Kuna Amerinds of Panama. Nephron. 1999;82:131–138.
9. Herrera M, Garvin JL. Recent advances in the regulation of nitric oxide in the kidney. Hypertension. 2005;45:1062–1067.
10. Chevaux KA, Jackson L, Villar ME, et al. Proximate mineral and procyandin content of certain foods and beverages consumed by Kuna Amerinds of Panama. J Food Composit Anal. 2001;14:553–563.
11. Fisher NDL, Hollenberg NKH. Flavanols for cardiovascular health: the science behind the sweetness. J Hypertension. 2005;23:1453–1459.
12. Fisher NDL, Hughes M, Gerhard-Herman M, et al. Flavanol-rich cocoa induces nitric-oxide-dependent vasodilation in healthy humans. J Hypertension. 2003;21:2281–2286.
13. Redman CW, Sargent IL. Latest advances in understanding preeclampsia. Science. 2005;308:1592–1594.
14. Takase B, Goto T, Hamabe A, et al. Flow-mediated dilation in brachial artery in the second half of pregnancy and pre-eclampsia. J Hum Hypertension. 2003;17:697–704.
15. Savvidou MD, Hingorani AD, Tsikas D, et al. Endothelial dysfunction and raised plasma concentrations of asymmetric dimethylarginine in pregnant women who subsequently develop pre-eclampsia. Lancet. 2003;361:1511–1517.
16. Schramm DD, Wang JF, Holt RR, et al. Chocolate procyanidins decrease the leukotriene-prostacyclin ratio in humans and human aortic endothelial cells. Am J Clin Nutr. 2001;73:36–40.
17. Karim M, McCormick K, Kappagoda CT. Effects of cocoa extracts on endothelium-dependent relaxation. J Nutr. 2000;130:2105S–2108S.

flavonoids; flavanols; cocoa; endothelial function; nitric oxide; hypertension

© 2006 Lippincott Williams & Wilkins, Inc.