NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Planning Committee for a Workshop on Potential Health Hazards Associated with Consumption of Caffeine in Food and Dietary Supplements; Food and Nutrition Board; Board on Health Sciences Policy; Institute of Medicine. Caffeine in Food and Dietary Supplements: Examining Safety: Workshop Summary. Washington (DC): National Academies Press (US); 2014 Apr 23.

Cover of Caffeine in Food and Dietary Supplements: Examining Safety

Caffeine in Food and Dietary Supplements: Examining Safety: Workshop Summary.

Show details

5Caffeine Effects on the Cardiovascular System

Much of the concern about caffeinated food and beverages and their potential health effects in vulnerable populations stems from several recent sudden cardiac deaths in adolescents being attributed to consumption of caffeinated energy drinks. However, during the workshop, some experts questioned the causal nature of the relationship. Others warned that, at the very least, the deaths are an early safety signal that warrants further investigation. Some workshop participants who spoke urged that until such investigation demonstrates the safety of caffeinated energy drinks in children, adolescents, pregnant women, caffeine-sensitive individuals, and other vulnerable populations, it would be prudent to restrict their use. In the Day 1, Session 3, panel, moderated by Stephen R. Daniels, M.D., Ph.D., Department of Pediatrics, University of Colorado School of Medicine, Denver, panelists explored the current state of the science on the effects of caffeine on the cardiovascular system. Box 5-1 describes the key points made by each speaker.

Box Icon

BOX 5-1

Key Points Made by Individual Speakers. John Higgins discussed data showing that endothelial cell function mediates the vascular effects of caffeine exposure, with implications for cardiac health. Caffeine in an individual at rest appears to improve endothelial (more...)


Presented by John P. Higgins, M.D., M.B.A., University of Texas Medical School

Endothelial cell function (ECF) serves an important role in mediating the vascular effects of caffeine exposure, according to John Higgins. He described normal1 and abnormal2 ECF and potential implications of abnormal ECF for cardiac health; explained how caffeine in individuals at rest appears to improve ECF but that caffeine in individuals during exercise appears to reduce ECF; and presented data suggesting that energy drinks in individuals at rest also reduce ECF.

Endothelial Cell Function

Endothelial cells form the inner lining of blood vessels and serve both basal and inducible metabolic and synthetic functions (Sumpio et al., 2002). Among other multiple tasks, normal ECF serves an important role in regulating vascular tone (i.e., blood vessel tone), preventing thrombosis (i.e., the ability of blood to clot in the artery), and preventing arterial damage by acting as a barrier. Higgins described ECF as a “balancing act,” with normal ECF being associated with vasodilatation (i.e., larger arteries), thromboresistance (i.e., thinner blood, which prevents blood clots), and antiadhesion. With respect to antiadhesion, Higgins compared normal ECF to the Teflon coating on a frying pan: when it is working well, things do not stick. The molecules that appear to be important for normal ECF are nitric oxide, prostaglandin I2, endothelium-derived hyperpolarizing factor, and bradykinin.

Abnormal ECF, on the other hand, manifests as vasoconstriction (i.e., smaller arteries), procoagulant effects (i.e., blood clot), and proadhesion, said Higgins. Molecules that appear to play an important role in abnormal ECF include renin, angiotensin, endothelin 1, and others.

Abnormal ECF is important in both the short term and the long term. In the short term, during stress or certain exposures—for example, in cold temperatures or during exposure to cigarette smoke or cocaine—abnormal ECF impairs the ability of arteries to dilate normally and potentially could result in a supply-demand imbalance, that is, with the heart beating harder and needing more blood flow while at the same time not being able to open up the arteries to improve blood flow. This supply-demand imbalance could in the short term lead to ischemia and possibly cardiac arrhythmia. In the long term, abnormal ECF can lead to hypertension, atherosclerosis, cardiovascular disease, coronary disease, and peripheral artery disease.

Improving ECF is a desirable goal, with many ways to do so, including exercise, smoking cessation, certain antioxidants (e.g., vitamin C, flavonoids in dark chocolate), cholesterol-lowering statins, omega-3 fatty acids, glycemic control in diabetes, L-arginine (i.e., a precursor to nitric oxide), and angiotensin-converting enzyme inhibitors and angiotensin-receptor blockers (Widlansky et al., 2003; Lee et al., 2012).

Caffeine in Individuals at Rest Improves ECF

Caffeine in individuals at rest is believed to improve ECF by increasing intracellular calcium, which in turn stimulates expression of endothelial nitric oxide synthase, which itself stimulates the endothelial cells to produce nitric oxide. The nitric oxide then diffuses to the vascular smooth muscle, which lies just underneath the endothelial cells, and results in vascular smooth muscle vasodilatation (Echeverri et al., 2010). Caffeine can also bind directly to the vascular smooth muscle cell receptors and, through similar mechanisms, cause vasodilatation (Echeverri et al., 2010).

Higgins described two in vivo3 studies on the ECF effects of caffeine in individuals at rest (Higgins and Babu, 2013). The first study involved 40 individuals, 33 of whom were male, with an average age of 53 years, all of whom consumed 200 mg of caffeine and were tested 60 minutes later using flow-mediated dilatation of the brachial artery. The researchers found that resting flow-mediated dilatation increased 10 percent after caffeine ingestion (p < 0.001). The second study involved 10 individuals, all males, with an average age of 27, all of whom consumed 300 mg of caffeine and were tested 60 minutes later using a strain-gauge plethysmograph to measure forearm blood flow. The researchers also measured blood flow responses to acetylcholine, which is an endothelium-dependent vasodilator, and to sodium nitroprusside, which is an endothelium-independent vasodilator. They found that resting forearm blood flow was not affected by caffeine but that resting forearm blood flow response to acetylcholine increased by 25 percent (p < 0.05). According to Higgins, these latter results suggest that the endothelium is very important in the vasorelaxation effect of caffeine in individuals at rest.

Equally significant, in Higgins's opinion, caffeine blocks adenosine receptors, which are important in dilating coronary arteries to augment coronary blood flow during exercise (Echeverri et al., 2010). This finding is important, he explained, because adenosine receptors are present throughout circulation where, in most places, they vasodilate, that is, they make arteries larger and thereby increase blood flow. For example, in the coronary arteries, the adenosine 2a receptor results in vasodilation. In the aorta, the adenosine 2b receptor does. Caffeine competitively blocks those and all other adenosine receptors, resulting in a compensatory increase in adenosine by the body, which in turn stimulates circulating chemoreceptors and other receptors, leading to an increase in sympathetic tone, catecholamines, peripheral vascular resistance, and renin secretion. These effects manifest as increased blood pressure, with systolic blood pressure increasing by 7 mm and diastolic blood pressure by 3 mm 60 minutes after ingestion of 300 mg of caffeine.

Caffeine Plus Exercise Decreases ECF

As a sports cardiologist, Higgins is especially interested in ECF as measured by myocardial blood flow. He described results from three studies based on measurements of either myocardial perfusion by positron emission tomography or brachial artery ECF by flow-mediated dilation (Higgins and Babu, 2013). Flow-mediated dilation measurements are also an accepted surrogate for coronary artery ECF (i.e., how well the brachial arteries in the arms can dilate correlates with how well the coronary arteries can dilate).

Higgins described the first study, which involved 15 subjects, 5 of whom were male, with an average age of 58 years, all of whom received 200 mg of pure caffeine and were then tested 50 minutes later while bicycling. Positron emission tomography was used to measure myocardial perfusion. The researchers found that exercise-induced myocardial blood flow response decreased 14 percent after caffeine ingestion (p < 0.05). So caffeine ingestion followed by exercise on a bicycle reduced coronary blood flow. The second study involved 18 individuals, 11 of whom were male, with an average age of 27 years, all of whom received 200 mg of pure caffeine and were then tested 50 minutes later while bicycling. The researchers found that exercise-induced myocardial blood flow response decreased 22 percent after caffeine ingestion (p < 0.01). The third study involved 10 individuals, with an average age of 30, who were administered 360 mg of caffeine and their forearm blood flow measured at baseline and then again every 20 minutes during 55 minutes of bicycling. Forearm blood flow was measured using a venous plethysmographic exclusion technique with a wrist cuff method of flow-mediated dilatation. The researchers found that, with individuals at rest, caffeine had no effect on forearm blood flow. During exercise, however, caffeine attenuated the usual increase in forearm blood flow by 53 percent (p < 0.05). In sum, caffeine plus exercise appears to decrease blood flow.

Energy Drinks in Individuals at Rest Decrease ECF

Higgins expressed concern that children and teenagers can purchase caffeine-containing energy drinks in stores. He observed that it is not uncommon for today's youth to consume cans of energy drinks at soccer game half-times, instead of the cut-up oranges or bananas that he and his peers used to consume as children during their soccer games/school events. He is concerned because he and his colleagues have witnessed many emergency room visits by children and adolescents after having consumed energy drinks. He also mentioned the wrongful death lawsuit filed against an energy drink company in October 2012 and a March 2013 letter to the FDA commissioner asking the FDA to examine the case reports of sudden cardiac death associated with energy drink consumption.

With this concern in mind, Higgins described the results of two studies on energy drinks and ECF. First, Worthley et al. (2010) measured two types of ECF in 50 healthy volunteers, including 34 males, with an average age of 22 years. The researchers measured adenosine disphosphate–induced platelet aggregation and the reactive hyperemia index (i.e., how the artery is able to dilate) both before and 1 hour after the volunteers consumed a 250-ml can of sugar-free energy drink. They reported a significant (p < 0.007) 13.7 percent increase in platelet aggregation following energy drink consumption, compared to basically no change in the control, and a significant (p < 0.05) reduction in reactive hyperemia index, again compared to a nonsignificant difference from baseline in the control. Also, not unexpectedly, according to Higgins, the researchers reported an increase in blood pressure following the energy beverage consumption (p < 0.05).

The second study Higgins described was based on his own research at the University of Texas Health Science Center. The study, SHADE-ONE (Study of Heart Effects from Adults Drinking Energy Beverages: On Endothelial Function), was a pilot study involving Higgins himself (Higgins, 2013). Measurements were taken prior to Higgins consuming a 24-ounce energy drink and then 90 minutes after consumption. At rest, Higgins's artery dilated to 0.42 cm and then to 0.45 cm after he performed flow-mediated dilatation using the standard cuff occlusion method. Ninety minutes after he drank the 24-ounce energy drink, his artery dilated to 0.43 cm but then only to 0.44 cm with maximal flow-mediated dilatation.

As part of SHADE-ONE, Higgins also measured percent flow-mediated dilatation at 50 minutes as well as at baseline and 90 minutes and found that it decreased over time and was lowest at 90 minutes. According to Higgins, most people's caffeine levels peak between 45 and 60 minutes after consumption. He found it interesting that with this energy drink, which he noted has other important ingredients in addition to caffeine, the maximal effect was observed at 90 minutes. The finding suggests to Higgins that there may be something about energy drinks that makes them “different beasts” than coffee and other modes of caffeine delivery. For example, maybe there is something in them that affects the pharmacokinetics or dynamics of caffeine by interacting with the caffeine and thereby having more deleterious effects on ECF.

Conclusions About Vascular Effects of Caffeine

Higgins concluded that in healthy individuals aged 22 to 59 years who consume 200 to 300 mg of caffeine, indirect tests indicate improved ECF and vasodilation at rest. So adults consuming this much caffeine during activities of daily living are likely safe, provided they are not caffeine sensitive, pregnant, or taking medication that interacts with caffeine or do not have a medical condition that is worsened by caffeine. For those who consume caffeine immediately before or during exercise, however, there could be harmful results. It appears that caffeine may attenuate the normal physiological mechanisms that help increase myocardial blood flow that occur during the increased demand of exercise. Researchers know that caffeine blocks adenosine receptors, thus reducing the ability of the coronary arteries to improve their flow commensurate with the increased myocardial demand of exercise. This could perhaps result in supply-demand ischemia. In healthy individuals aged 21 to 71 years who consume 200 to 300 mg of caffeine and then perform aerobic exercise 1 hour later, indirect tests indicate reduced ECF as measured by reduced myocardial blood flow. Finally, in healthy individuals aged 20 to 47 years who consume energy drinks, indirect tests indicate reduced ECF at rest.

Higgins called for more research on the effects of caffeine and energy drinks on ECF and the mechanisms underlying those effects and for more research on the safety of high-dose caffeine and energy drinks in younger individuals, caffeine-naïve individuals, and individuals who exercise 1 to 2 hours after consumption. In the 6 cases that he was aware of in which deaths were associated with energy drink consumption, affected individuals were between 12 and 19 years of age. He identified that age group as a potentially vulnerable population.


Presented by Jeffrey Goldberger, M.D., Northwestern University

A 1994 survey of several hundred physicians from Minnesota and Vermont found that 94 percent of those surveyed recommended reducing or stopping caffeine for patients complaining of palpitations (Hughes et al., 1988). Jeffrey Goldberger described the finding as “remarkable” and considered it his task for the workshop to examine whether the evidence supports that recommendation. In his experience, it is not often that 94 percent of physicians agree on something even when its benefits have been demonstrated in randomized clinical trials, such as the use of beta-blockers after myocardial infarction or anticoagulants for atrial fibrillation. It is difficult to get that kind of consensus and interesting to consider where it comes from.

There are many data sources on the effects of caffeine on arrhythmias, including case reports, animal studies, human physiologic studies, human small-case series, and human observational trials. The predominant focus of Goldberger's presentation was on human observational trials. He noted that most of the data comes from coffee-intake studies and emphasized the need to keep in mind, while reviewing these studies, the variation in the amount of caffeine in different coffee drinks.

There are also many end points to consider when evaluating the effects of caffeine on the heart, including physiologic surrogates (i.e., electrophysiological effects such as QRS duration, which is the time required to depolarize the ventricles), specific arrhythmias, and epidemiological outcomes. He focused on three types of specific arrhythmias: (1) atrial fibrillation, a common arrhythmia in middle-aged and older individuals; (2) premature ventricular complexes, which are extra beats that arise from the ventricles and are common at all ages and in people either with or without heart disease; and (3) arrhythmias that can lead to sudden cardiac death, which include ventricular fibrillation and very rapid ventricular tachycardia.

Animal Studies on the Effect of Caffeine on Cardiac Arrhythmias

Goldberger described two animal studies that he thought were especially interesting. First, Bellet et al. (1972) examined the effect of caffeine on ventricular fibrillation thresholds in dogs. The researchers measured the amount of energy required to induce ventricular fibrillation with shocks to the heart among both control dogs and dogs that had experienced myocardial infarctions. They found that having had a myocardial infarction reduced the ventricular fibrillation threshold. In both groups of dogs, the ventricular fibrillation threshold was reduced even further when caffeine was administered. Goldberger noted, however, that the caffeine dose used was 25 mg per kg, which would amount to about 1.75 gm in a 70-kg man, not a typical human dose. In a second animal study, Rashid et al. (2006) examined the effect of caffeine on the inducibility of atrial fibrillation. They found that caffeine actually decreased the window of vulnerability for atrial fibrillation. Goldberger remarked that he was unsure of the clinical implications of this finding but found the study interesting because it reflects the range of caffeine effects observed in animals in relation to cardiac arrhythmias.

Human Studies on the Effect of Caffeine on Arrhythmias and Other Cardiac End Points

Many health effects that are observed in association with caffeine exposure are those that occur on sympathetic excitation, according to Goldberger. Corti et al. (2002) examined the effect of coffee on sympathetic nerve activity in a placebo-controlled trial of 15 healthy volunteers (6 habitual and 9 nonhabitual coffee drinkers). A number of interventions were tested, including intravenous caffeine (250 mg) versus placebo and a triple espresso (which was designed to mimic the 250-mg intravenous treatment) versus a decaf triple espresso (nonhabitual coffee drinkers only). The researchers reported a sustained increase in caffeine levels in the intravenous caffeine group and, not surprisingly in Goldberger's opinion, a small increase in blood pressure and a drop in heart rate. Sympathetic nerve activity, as measured by a number of different techniques, also increased. The placebo group showed no change over time. With coffee drinking, there was no difference between habitual versus nonhabitual coffee drinkers with respect to sympathetic nerve activity or caffeine levels. A striking finding, in Goldberger's opinion, was that decaf administered to nonhabitual users increased blood pressure and sympathetic nerve activity. Habitual users showed no increase in blood pressure.

In another study, Jackman et al. (1996) examined caffeine effects on catecholamines in 14 athletes during intense exercise. The researchers orally administered 6 mg of caffeine per kg 1 hour before exercise. The exercise protocol involved cycling at 2 minutes at a power required to achieve maximum oxygen uptake, resting for minutes, cycling again at the same power for 2 minutes, resting for 6, and then cycling at the same power to exhaustion. They found a slight increase in exercise endurance and significantly higher plasma epinephrine levels at peak exercise in the caffeine group. In Goldberger's opinion, the findings serve as evidence of sympathoexcitation.

In one of the first studies on caffeine's effect on arrhythmias, Myers and Harris (1990) examined 35 patients with recent myocardial infarctions (within 5 to 10 days) in a double-blind crossover design. Patients received either 300 mg of caffeine plus an additional 150 mg of caffeine 4 hours later or a placebo. The researchers monitored the patients for 8 hours and then counted premature ventricular complexes (PVCs). They found no statistically significant difference in the number of PVCs. Goldberger remarked that PVCs are highly variable in general, which has always been problematic for studies on PVCs.

In another early study on arrhythmias, Chelsky et al. (1990) examined 222 patients, 86 percent with coronary artery disease and all habitual coffee users. All the patients were being evaluated for some sort of ventricular arrhythmia: nonsustained ventricular tachycardia, ventricular tachycardia, or ventricular fibrillation. The researchers attempted to induce very rapid rhythms in the patients' hearts both before and after caffeine consumption and found no difference in inducibility of ventricular arrhythmia on the basis of caffeine.

In what Goldberger described as a “curious” study of 600 patients who had experienced their first episode of atrial fibrillation, Mattioli et al. (2011) retrospectively examined caffeine intake in the days before the atrial fibrillation compared to usual intake. The researchers found that patients with moderately heavier caffeine use in the days prior to atrial fibrillation compared to usual intake had the lowest rate of successful spontaneous conversion to sinus rhythm. Patients with the lowest and highest intakes prior to atrial fibrillation had higher rates. The researchers called the response a U-shaped response.

In a 2011 review, Goldberger and Dan Pelchovitz (Pelchovitz and Goldberger, 2011) listed by size the studies they could find on the effect of caffeine on arrhythmia. Of those studies, three involved several thousand patients. Many more included far fewer numbers of patients. Goldberger highlighted one of the larger studies, the de Vreede-Swagemakers et al. (1999) study, which reported that, in a population with coronary artery disease (i.e., all the patients in the study had a clinical history of coronary artery disease), ingesting more than 10 cups of coffee per day was associated with an odds ratio of 55.7 for sudden cardiac death. This case-control study investigated 117 cases of sudden cardiac arrest and 144 controls, with controls matched by age and gender. A challenge for the investigators was to determine how many cups of coffee had been consumed by the patients who had died from sudden cardiac arrest. To answer that challenge, they asked the patients' relatives. Thus, they also asked relatives of the controls how much coffee the controls had consumed. The researchers found dramatically fewer individuals in the control group who had consumed more than 10 cups of coffee per day, compared to individuals in the sudden cardiac arrest group. Interestingly, in Goldberger's opinion, what are typically considered risk factors, that is, hypercholesterolemia, diabetes mellitus, and smoking, were not identified as risk factors. The question for Goldberger is how to interpret such data given that they conflict with otherwise overwhelming data from other studies.

Since the Pelchovitz and Goldberger (2011) review, three additional observational studies involving thousands of patients have been published. Among all the observational studies conducted thus far, a total of about 315,000 patients have been evaluated. The average caffeine doses among these studies are consistent with information presented earlier during the workshop about amounts of caffeine intake (i.e., 2 cups of coffee daily: 148 mg/day for men, 285 mg/day for women; 5 cups daily: 274 mg/day for men and 232 mg/day for women), except a study conducted in Europe where average intake was greater (584 mg/day). Peak doses tend to be in 600 to 800 mg/day in the United States and greater than 1 g/day in Europe. According to Goldberger, the preponderance of these studies showed no increased risk of arrhythmias as a result of caffeine consumption.

The largest of these population studies, Klatsky et al. (2011), showed an inverse relationship between coffee intake and risk of hospitalization for arrhythmias, with an average follow-up of 17.6 years. That study was based on a Kaiser Permanente database of patients admitted to the hospital for arrhythmias. For “any arrhythmia,” the odds ratio was 0.97 per cup per day, which represented a statistically significant decline as the number of cups of coffee per day increased. Odds ratios for several other diagnoses were either not significant or borderline significant. For premature beats, again the odds ratio, 0.87, represented a statistically significant decline in risk as coffee consumption increased. Whether the declines observed in Klatsky et al. (2011) are “real” is hard to know, in Goldberger's opinion. There are several caveats to population studies. Individuals who are sensitive to caffeine likely do not consume; there are ascertainment issues; exposure changes over time; source of caffeine varies and may matter, with other additives potentially having other effects; and most data do not include adolescents.

More broadly, there are studies that have examined other cardiovascular outcomes, some of which have shown negative outcomes in association with caffeine use and others positive outcomes. For example, Lopez-Garcia et al. (2006), a health professional's follow-up study, reported lower rates of coronary heart disease with caffeine consumption. But other studies have shown the opposite, according to Goldberger.

Conclusions About Caffeine and Arrhythmia Risk

Goldberger concluded by describing one final study, Graboys et al. (1989), of 50 individuals, all with significant arrhythmias and structural heart disease. The individuals were administered caffeine on one day and no caffeine on a successive day. The researchers found that caffeine and catecholamine levels increased with caffeine consumption, as expected, but they observed no change in PVCs with caffeine consumption. The researchers concluded, “Although patients with cardiac disease are frequently warned about the potential harmful effects of caffeine, this clinical advice is based primarily on anecdote and folklore” (p. 639). That was 25 years ago. In Goldberger's opinion, some data today suggest that caffeine effects are present, but the prevailing evidence shows no increase in arrhythmia. Moreover, what effects do exist are dose dependent and different in habitual versus nonhabitual users. Researchers have demonstrated mild changes in hemodynamic parameters (heart rate and blood pressure), a slight increase in sympathetic activity, and small changes in cardiac electrophysiologic properties.


Presented by Ahmed El-Sohemy, Ph.D., University of Toronto

In Ahmed El-Sohemy's opinion, the marketing of energy drinks to children and adolescents is a major issue. Isolated case reports of premature death following consumption of an energy drink usually occur in the context of some kind of physical exertion or activity and primarily among youth and adolescents. Despite denials by some manufacturers, El-Sohemy opined, it is difficult to argue that such drinks are not intended to appeal to youth. For example, recently at a high school in Winnipeg, Canada, an energy drink company was distributing certificates containing advertising messages to graduating students. Many of those students were under the age of 18.

El-Sohemy showed some slides courtesy of Jim Shepherd, whose son died a few years ago at a paintball event. Although the finding was not conclusive, there was reason to believe that consumption of an energy drink at that event, which was the first time Shepherd's son had consumed one, may have triggered the fatal cardiac event. The slides contained images of some of the products and ways that El-Sohemy believes they are being marketed to youth. For example, a product that El-Sohemy said was taken off the shelves had a label at the top of the can that read, “The Legal Alternative.” That sort of marketing appeals to risk-taking behavior, in El-Sohemy's opinion. Other types of caffeinated products on the shelves are of concern, too, El-Sohemy said. These products include brownies, gummy bears, potato chips, and gum. The Lancet published a case report several years ago describing the hospitalization of a 13-year-old boy with tachycardia and elevated blood pressure (Natale et al., 2009). The boy had consumed 2 packs of gum containing 160 mg of caffeine. There was good evidence, according to El-Sohemy, that consumption of the gum was responsible for the cardiovascular event that required hospitalization.

Globally, coffee is “still king” with respect to caffeine exposure, El-Sohemy said. Coffee is the second most widely traded commodity, after oil. Among adults in many parts of the world, coffee is still the biggest source of caffeine. Of course, El-Sohemy observed, a cup of coffee is no longer “just a cup of coffee.” For example, Tim Hortons (a restaurant chain in Canada and the United States) recently added a new extra-large-sized coffee, a 24-ounce cup containing several hundred milligrams of caffeine. El-Sohemy pointed this out because many energy drink manufacturers have argued that if energy drinks are going to be regulated, then coffee should be regulated too because many cans of energy drinks have an equivalent amount of caffeine as a cup of coffee. In fact, they argue that some cups of coffee have more caffeine than what is found in energy drinks. However, it is important to note, El-Sohemy said, that a distinction exists between caffeine from energy drinks and caffeine from coffee. Caffeine is caffeine from a chemical structure perspective, but there is a big difference in terms of peak concentrations of caffeine between slowly sipping a hot beverage versus chugging a cold beverage.

Is Coffee Associated with Cardiovascular Disease

Dozens of studies have examined the association between coffee and cardiovascular disease. Not surprisingly, in El-Sohemy's opinion, some have shown an increased risk of cardiovascular disease with caffeine exposure, others have shown no effect, and yet others have shown a decreased effect with moderate consumption. Some studies have shown a U-shaped or J-shaped association, with moderate consumption associated with the lowest risk. There are many possible reasons for these inconsistencies. One is the genetic background of the population being studied. El-Sohemy and his research team wanted to explore the possibility that individuals with certain genotypes are more vulnerable and at greater risk while individuals with certain other genotypes experience no effect or might actually benefit from moderate consumption.

El-Sohemy explained that he and his team were interested specifically in caffeine. After all, he said, coffee is a complex beverage with many kinds of bioactive substances. Some, such as the polyphenols, with their antioxidant properties, are believed to have beneficial effects, whereas others, such as the diterpenoids, which are known to raise low-density lipoprotein cholesterol, could have adverse effects.

Caffeine is broken down almost exclusively by the drug-metabolizing liver enzyme CYP1A2 and converted into a more water-soluble compound, paraxanthine, which itself is rapidly broken down into other water-soluble compounds. CYP1A2 catalyzes the rate-limiting detoxification of caffeine, with the gene that codes for CYP1A2 having a common polymorphism (–163 A/C) with a profound effect on enzyme activity. Carriers of the C allele are slow metabolizers. Individuals homozygous for the A allele have a fourfold higher rate of caffeine metabolism. El-Sohemy's team reasoned that if caffeine is a component in coffee that could increase the risk of heart attack, then slow metabolizers should be at higher risk than fast metabolizers because caffeine lingers longer in slow metabolizers' systems.

As described in Cornelis et al. (2006), El-Sohemy and colleagues examined genetic variation in CYP1A2 and coffee intake in more than two thousand cases of a first acute myocardial infarction and an equal number of controls matched for age, sex, and area of residence. They used a food frequency questionnaire to assess coffee consumption and other sources of caffeine. They found that 90 percent of caffeine intake came from coffee. They genotyped participants from fasting blood samples. Without taking genetics into account, but taking into account potential confounding factors such as smoking, physical activity, and saturated fat intake, the researchers found that consuming four or more cups of coffee a day was associated with about a 36 percent increased risk of a myocardial infarction and a statistically significant odds ratio of 1.36.

El-Sohemy explained that, if they had stopped their research with that finding, they would have concluded that drinking four or more cups of coffee per day is associated with an increased risk of a heart attack. But the real question they wanted to address was related to caffeine intake and its association with myocardial infarction. Again, if caffeine increases the risk, then one would expect slow metabolizers to be at a higher risk. Indeed, that is what they found when they stratified the study population by CYP1A2 genotype (see Figure 5-1). Among slow metabolizers, two to three cups of coffee per day was associated with a significantly increased risk. Among the fast metabolizers, there was no increased risk. If anything, the data show signs of a U-shaped association, according to El-Sohemy.

FIGURE 5-1. Odds ratios of risk of myocardial infarction with coffee intake.


Odds ratios of risk of myocardial infarction with coffee intake. NOTES: AA = fast caffeine metabolizer, AC + CC = slow caffeine metabolizer. Top panel: all study participants; bottom panel: study participants less than 50 years of age. SOURCE: Cornelis (more...)

At the time those data were collected, researchers believed that that caffeine more likely acted as a trigger of cardiovascular disease among younger individuals in particular. As expected, Cornelis et al. (2006) observed a shift in odds ratios for subjects younger than 50 years of age (see Figure 5-1), with slow metabolizers showing a more pronounced risk of heart attack, almost a fourfold increased risk, and with a significant protective effect of moderate consumption among fast metabolizers. El-Sohemy suspects that, at moderate levels of consumption, fast metabolizers are able to efficiently eliminate the caffeine, which would otherwise be masking some of the beneficial effects of the polyphenols and other bioactive substances. But once these individuals reach four or more cups a day, even though they are fast metabolizers, their CYP1A2 enzyme begins to become saturated so that the adverse effects of caffeine begin to counter the beneficial effects of the polyphenols and other compounds.

El-Sohemy reiterated that the only difference between slow and fast metabolizers who drink four or more cups a day of coffee in Cornelis et al. (2006) is a single nucleotide polymorphism that affects the rate at which caffeine is eliminated from the body. Because caffeine is the only major substance in coffee that is known to be detoxified by CYP1A2, the findings strongly implicate caffeine as a trigger for the increased risk of heart attack.

As El-Sohemy noted, that study attracted a lot of media attention, with headlines reading, “Why two cups of coffee can damage your heart” and “Gene that could make your next coffee your last.” But it wasn't until a few years later that another research group built on the findings by examining whether CYP1A2 might also explain some of the inconsistencies in studies linking coffee to risk of hypertension (Palatini et al., 2009). They essentially achieved the same results, with an increased risk of hypertension as coffee consumption increases among slow metabolizers but with a decreased risk of hypertension among fast metabolizers. Again, if genetics were not taken into account, one would conclude that coffee has no effect on hypertension. The study was prospective. The researchers examined prehypertensive individuals, genotyped them, and followed them. They also investigated the relationship between CYP1A2 genotype and catecholamines and, again, found that epinephrine concentrations increased with increased coffee consumption only among the slow metabolizers.

In another example of the importance of genetic variation in understanding the cardiovascular risks associated with caffeine consumption, a research group in Finland reported that genetic variation in the catechol-O-methyltransferase (COMT) gene, which is involved in the metabolism of catecholamines such as epinephrine, is associated with varying risk of acute coronary events (Happonen et al., 2006). Specifically, they found that individuals with the genotype associated with high COMT activity showed no increased risk with increased coffee consumption but that individuals with the genotype associated with low COMT activity did show an increased risk with increased coffee consumption. More recently, Brathwaite et al. (2011) reported that COMT could also explain why some people experience increased heart rate following caffeine consumption.

Personalized Dietary Advice Versus Public Health Recommendations

El-Sohemy concluded by emphasizing the importance of individual variation and the challenge of reconciling public health advice with personalized dietary advice. A “one-size-fits-all” approach clearly does not apply when it comes to caffeine consumption. There are probably many other genetic variants, in addition to those described here, that explain other types of responses. In El-Sohemy's opinion, it is highly unlikely that the several cases of premature death following the consumption of energy drinks are caused by CYP1A2 variation, which is very common in the population, or COMT variation, which is also fairly common. But it is likely that the cases are caused by some other genetic variant. He said, “But what we don't know is how common that other genetic variant is because it has not yet been identified.” In terms of regulation, he stressed the importance of taking into account these vulnerable (genetic) subgroups and ensuring that they are protected.


Following El-Sohemy's presentation, workshop participants were invited to ask questions of the three panelists. Most of the discussion revolved around the future research needs on the cardiovascular effects of caffeine exposure, including in vulnerable populations, and differences between caffeine in energy drinks versus coffee.

Future Research Needs

Some workshop participants expressed disagreement regarding the urgency of concern with consumption of energy drinks, with one audience member remarking that he has not seen “in the real clinical world” the adverse effects being studied in the lab. He suggested that if a problem with energy drink consumption did exist, more people would be admitted for arrhythmias, heart attacks, and so forth. In response, Higgins explained that, as some workshop presenters emphasized, not all individuals are equal. It may be that the cardiologists who treat adults are not seeing the same problems that the cardiologists who treat children are seeing. Some people may be vulnerable by age, others by exposure to caffeine, and still others because of a genetic predisposition. Also, the substrates are different. In his opinion, coffee, with its many antioxidants and other components, is not the same as pure caffeine. Nor are either coffee or pure caffeine equal to caffeine-containing energy drinks. He said that “obviously more research is needed” with respect to why some individuals have greater reactions than others.

Goldberger added that, although there are no large differences in arrhythmia on a population level, nonetheless he occasionally comes across patients with arrhythmias who report sensitivity to caffeine. In addition to genetic factors, other factors may to help identify the small subset of individuals who are more susceptible to arrhythmias on exposure to caffeine.

Goldberger was asked whether any of the studies he described involved individuals under the age of 18 and whether any of the studies differentiated between caffeine versus coffee versus energy drinks versus other energy products. Goldberger replied that the preponderance of population studies have examined coffee intake, with not much additional information available. Most have not addressed vulnerable populations.

Implications of a Study with an “N” of One

One audience member expressed concern that a sample size of 1 (i.e., Higgins's SHADE-ONE study) does not reflect “true variation” and stressed the importance of replication before drawing generalizations. He wondered what the ECF response would be in a large population and how ECF would vary in response to different doses. Higgins cited Worthley et al. (2010), a study with an N of 50, which presented clear evidence of endothelial dysfunction as measured by reactive hyperemia and platelet aggregation. He remarked that there are other larger (i.e., larger than N = 1) studies as well and that the pilot study he mentioned with an N of one (Higgins, 2013) is currently being followed up with a larger study with an N of about 50.

Different Studies Have Different End Points

Yet another audience member observed that Higgins's research “flies in the face of decades of research showing that caffeine actually increases performance.” Higgins explained that his research on ECF does not relate to performance. He did not disagree with the audience member's claim about other studies on performance, but he has been looking only at arterial function.

Differences Between Caffeine in Energy Drinks Versus Coffee

One audience member disputed El-Sohemy's claim that peak concentrations from chugging a cold drink are different from those from slowly sipping a hot beverage. According to the audience member, the dose response is the same for similar doses regardless of whether the dose is being consumed slowly or quickly. El-Sohemy replied that the actual dose, the actual amount of caffeine being consumed, is not the only issue. In terms of peak plasma concentrations, which can have important physiological effects, chugging a cold beverage leads to a higher peak plasma concentration and could have a profound effect even if the dose is smaller than the dose in a hot beverage sipped slowly.


  • Bellet S, Horstmann E, Roman LR, DeGuzman NT, Kostis JB. Effect of caffeine on the ventricular fibrillation threshold in normal dogs and dogs with acute myocardial infarction. American Heart Journal. 1972;84:215–227. [PubMed: 5075075]
  • Brathwaite JM, Da Costa LA, El-Sohemy A. Catechol-o-methyltransferase genotype is associated with self-reported increased heart rate following caffeine consumption. Journal of Caffeine Research. 2011;1:123–130.
  • Chelsky LB, Cutler JE, Griffith K, Kron J, McClelland JH, McAnulty JH. Caffeine and ventricular arrhythmias: An electrophysiologic approach. Journal of the American Medical Association. 1990;264:2236–2240. [PubMed: 2214101]
  • Cornelis MC, El-Sohemy A, Kabagambe EK, Campos H. Coffee, CYP1A2 genotype, and risk of myocardial infarction. Journal of the American Medical Association. 2006;295:1135–1141. [PubMed: 16522833]
  • Corti R, Binggeli C, Sudano I, Spieker L, Hänseler E, Ruschitzka F, Chaplin WF, Lüschers TF, Noll G. Coffee acutely increases sympathetic nerve activity and blood pressure independently of caffeine content: Role of habitual versus nonhabitual drinking. Circulation. 2002;106:2935–2940. [PubMed: 12460875]
  • De Vreede-Swagemakers JJ, Gorgels AP, Weijenberg MP, Dubois-Arbouw WI, Golombeck B, van Ree JW, Knottnerus A, Wellens HJ. Risk indicators for out-of-hospital cardiac arrest in patients with coronary artery disease. Journal of Clinical Epidemiology. 1999;52:601–607. [PubMed: 10391652]
  • Echeverri D, Montes FR, Cabrera M, Galán A, Prieto A. Caffeine's vascular mechanisms of action. International Journal of Vascular Medicine. 2010 PMCID: PMC3003984. article ID: 834060. doi: 10.1155/2010/834060. [PMC free article: PMC3003984] [PubMed: 21188209]
  • Graboys TB, Blatt CM, Lown B. The effect of caffeine on ventricular ectopic activity in patients with malignant ventricular arrhythmia. Archives of Internal Medicine. 1989;149:637–639. [PubMed: 2465748]
  • Happonen P, Voutilainen S, Tuomainen TP, Salonen JT. Catechol-o-methyltransferase gene polymorphism modifies the effect of coffee intake on incidence of acute coronary events. PLoS ONE. 2006;1(e117) [PMC free article: PMC1762420] [PubMed: 17205121]
  • Higgins JP. Endothelial function acutely worse after drinking energy beverage. International Journal of Cardiology. 2013 http://www​.sciencedirect​.com/science/article​/pii/S0167527313011728 . [PubMed: 23876461]
  • Higgins JP, Babu KM. Caffeine reduces myocardial blood flow during exercise. American Journal of Medicine. 2013;126(8):730.e731–730.e738. [PubMed: 23764265]
  • Hughes JR, Amori G, Hatsukami DK. A survey of physician advice about caffeine. Journal of Substance Abuse. 1988;1:67–70. [PubMed: 2485281]
  • Jackman M, Wendling P, Friars D, Graham TE. Metabolic catecholamine, and endurance responses to caffeine during intense exercise. Journal of Applied Physiology. 1996;81:1658–1663. [PubMed: 8904583]
  • Klatsky AL, Hasan AS, Armstrong MA, Udaltsova N, Morton C. Coffee, caffeine, and risk of hospitalization for arrhythmias. Permanente Journal. 2011;15:19–25. [PMC free article: PMC3200095] [PubMed: 22058665]
  • Lee R, Channon KM, Antoniades C. Therapeutic strategies targeting endothelial function in humans: Clinical implications. Current Vascular Pharmacology. 2012;10:77–93. [PubMed: 22112349]
  • Lopez-Garcia E, van Dam RM, Willett WC, Rimm EB, Manson JE, Stampfer MJ, Rexrode KM, Hu FB. Coffee consumption and coronary heart disease in men and women: A prospective cohort study. Circulation. 2006;113:2045–2053. [PubMed: 16636169]
  • Mattioli AV, Farinetti A, Miloro C, Pedrazzi P, Mattioli G. Influence of coffee and caffeine consumption on atrial fibrillation in hypertensive patients. Nutrition, Metabolism, and Cardiovascular Diseases. 2011;21:412–417. [PubMed: 20167459]
  • Myers MG, Harris L. High dose caffeine and ventricular arrhythmias. Canadian Journal of Cardiology. 1990;6:95–98. [PubMed: 1692755]
  • Natale F, Cirillos C, Di Marco GM, di Vetta LS, Aronne L, Siciliano A, Mocerino R, Tedesco MA, Golino P, Calabrò R. When chewing gum is more than just a bad habit. Lancet. 2009;373(1918) [PubMed: 19482217]
  • Palatini P, Ceolotto G, Ragazzo F, Dorigatti F, Saladini F, Papparella I, Mos L, Zanata G, Santonastaso M. CYP1A2 genotype modifies the association between coffee intake and the risk of hypertension. Journal of Hypertension. 2009;27:1594–1601. [PubMed: 19451835]
  • Pelchovitz DJ, Goldberger JJ. Caffeine and cardiac arrhythmias: A review of the evidence. American Journal of Medicine. 2011;124:284–289. [PubMed: 21435415]
  • Rashid A, Hines M, Scherlag BJ, Yamanashi WS, Lovallo W. The effects of caffeine on the inducibility of atrial fibrillation. Journal of Electrocardiology. 2006;39:421–425. [PMC free article: PMC2257921] [PubMed: 16919674]
  • Sumpio BE, Riley JT, Dardik A. Cells in focus: Endothelial cell. International Journal of Biochemistry and Cell Biology. 2002;34:1508–1512. [PubMed: 12379270]
  • Widlansky ME, Gokce N, Keaney JF Jr, Vita JA. The clinical implications of endothelial dysfunction. Journal of the American College of Cardiology. 2003;42:1149–1160. [PubMed: 14522472]
  • Worthley MI, Prabhus A, De Sciscio P, Schultz C, Sanders P, Willoughby SR. Detrimental effects of energy drink consumption on platelet and endothelial function. American Journal of Medicine. 2010;123:184–187. [PubMed: 20103032]



Normal endothelial function is characterized by vasodilatation, thromboresistance, and blood cell antiadhesion.


Abnormal endothelial function is characterized by vasoconstriction of the arteries, procoagulant effects, and proadhesion of blood cells.


In vivo refers to experiments conducted in a living organism: plant, animal, or human.

Copyright 2014 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK202224


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.1M)

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...