, October 25, 2011, by David Cohen  —  The biofuel cell, uses glucose and oxygen at concentrations found in the body to generate electricity.

Plugging gadgets into a socket in the wall, or loading them with batteries – or maybe even unfurling a solar panel – is how most of us think of getting electricity. But what about plugging them into your body?

It may sound far fetched, but under the shadow of the Alps, Dr Serge Cosnier and his team at the Joseph Fourier University of Grenoble have built a device to do just that. Their gadget, called a biofuel cell, uses glucose and oxygen at concentrations found in the body to generate electricity.

Some ideas, some technologies may sound like science fiction, but they are fast becoming science fact. In our eight-part series we will be exploring ideas that are the future of technology.

They are the first group in the world to demonstrate their device working while implanted in a living animal. If all goes to plan, within a decade or two, biofuel cells may be used to power a range of medical implants, from sensors and drug delivery devices to entire artificial organs. All you’ll need to do to power them up is eat a candy bar, or drink a coke.

Biofuel cells could kick-start a revolution in artificial organs and prosthetics that would transform tens of thousands of lives every year.

A new range of artificial, electrically-powered organs are now under development, including hearts, kidneys, and bladder sphincter, and work has begun on fully-functioning artificial limbs such as hands, fingers, and even eyes. But they all have one Achilles heel: they need electricity to run.

Batteries are good enough for implants that don’t need much power, but they run out fast, and when it comes to implants, that is more than just an inconvenience, it is a fundamental limitation.

Even devices that do not use much power, such as pacemakers, have a fixed lifespan because they rely on batteries.

They usually need their power packs replaced 5 years after implantation. One study in the US found that one in five 70 year-olds implanted with a pacemaker, survived for another 20 years – meaning this group needed around 3 additional operations after the initial implant, just to replace the battery.

Each operation is accompanied by the risk of the complications of surgery, not something anybody should have to face if it is avoidable.

Other devices such as artificial kidneys, limbs or eyes, would have such high energy demands that users would have to change their power source every few weeks to keep them working. It is simply impractical to use batteries in these devices.

That is where biofuel cells come in. Dr Cosnier and his team are one of a growing number of researchers around the world developing the technology in an attempt to side-step this inherent limitation.

Bodily fluids

The fuel cells are made from a compressed push of enzymes and carbon nanotubes.

At heart, biofuel cells are incredibly simple. They are made of two special electrodes – one is endowed with the ability to remove electrons from glucose, the other with the ability to donate electrons to molecules of oxygen and hydrogen, producing water.

Pop these electrodes into a solution containing glucose and oxygen, and one will start to rip electrons off the glucose and the other will start dumping electrons onto oxygen. Connect the electrodes to a circuit and they produce a net flow of electrons from one electrode to the other via the circuit – resulting in an electrical current.

Glucose and oxygen are both freely available in the human body, so hypothetically, a biofuel cell could keep working indefinitely. “A battery consumes the energy stored in it, and when it’s finished, it’s finished. A biofuel cell in theory can work without limits because it consumes substances that come from physiological fluids, and are constantly being replenished,” said Dr Cosnier.

“A bio fuel cell in theory can work without limits because it consumes substances that come from physiological fluids,” said Dr Serge Cosnier of Joseph Fourier University

The idea of powering fuel cells using glucose and oxygen found in physiological fluids was first suggested in the 1970s, but fell by the wayside because the amount of energy early prototypes produced was too little to be of practical use.

However, in the 2002, advances in biotechnology spurred Itamar Willner, a researcher at the Hebrew University in Jerusalem, to dust down the idea and give it a fresh look.

In a paper published in the prestigious journal Science, he speculated that thanks to advances in biotechnology, the day would come when devices such as artificial limbs and organs would soon be powered by biofuel cells that create electricity from bodily fluids.

“Since then biofuel cells have received a huge amount of attention,” said Dr Eileen Yu, a researcher at Newcastle University, who is part of UK-wide multi-university project to develop biofuel cells.

Nano technology

The key to the recent breakthroughs has been our understanding of rather special biological molecules called enzymes. Enzymes are naturally occurring molecules that speed up chemical reactions. Researchers studying bio fuel cells have discovered that one particular enzyme, called glucose oxidase, is extremely good at removing electrons from glucose. “It is very efficient at generating electrons,” said Prof Willner.

Spurred by new developments in enzyme manipulation, and the growth in availability of carbon nanotubes – which are highly efficient electrical conductors – many groups around the world have developed bio fuel cells capable of producing electricity.

Dr Cosnier and his team decided to take things one step further. “In the last 10 years there has been an exponential increase in research, and some important breakthroughs in enzyme research,” he said.

He decided it was time to make the first attempt to take the cumulative knowledge of the last decade of research and engineer it into a device the size of a grain of rice that could generate electricity while implanted inside a rat.




Tiny bio fuel cells sit inside the body turning glucose and oxygen into power.

In 2010, they tested their fuel cell in a rat for 40 days and reported that it worked flawlessly, producing a steady electrical current throughout, with no noticeable side effects on the rat’s behaviour or physiology.

Their system is surprisingly straightforward. The electrodes are made by compressing a paste of carbon nanotubes mixed with glucose oxidase for one electrode, and glucose and polyphenol oxidase for the other.

The electrodes have a platinum wire inserted in them to carry the current to the circuit. Then the electrodes are wrapped in a special material that prevents any nanotubes or enzymes from escaping into the body.

Finally, the whole package is wrapped in a mesh that protects the electrodes from the body’s immune system, while still allowing the free flow of glucose and oxygen to the electrodes. The whole package is then implanted in the rat.

“It is an important step towards demonstrating the translation of basic research into a practical device,” said Willner. “It shows the feasibility of making an implantable package.”

Implantation in a rat was a good proof of concept, said Dr Cosnier, but it had drawbacks. “Rats are so small that the production of energy is insufficient to power a conventional device.”

Next he plans to scale up his fuel cell and implant it in a cow. “There is more space, so a larger fuel cell can be implanted, meaning a greater current will be generated.”

Dr Cosnier hopes it will be enough to power a transmitter that will be able to beam out of the cow information about the device and control sensors inside the animal.

More power



Fuel cells are wrapped in a mesh to prevent the body rejecting them.

There is still a long way to go. Prof Willner explains that, while the enzyme glucose oxidase has performed optimally, the efficiency of the electron-donating enzymes could still be dramatically improved. He is optimistic that breakthroughs will be made.

“Based on the current rate of progress, I am confident we will see exciting developments in the next decade,” said Prof Willner.

Dr Cosnier agrees that there is a lot of room for improvement. “Today we can generate enough power to supply an artificial urinary sphincter, or pacemaker. We are already working on a system that can produce 50 times that amount of power, then we will have enough to supply much more demanding devices,” he said.

Implants aren’t the only place you may find bio fuel cells in the future. The electronics giant Sony recently announced that it had created a biofuel cell fuelled with glucose and water that was capable of powering an MP3 player. “In 10 years time you may see bio fuel cells in laptops and mobile phones,” said Prof Willner.

Dr Cosnier points out that bio fuel cells would be especially useful in places where there is no electricity supply to recharge your batteries. “If you were in a country without electricity, and needed to re-charge a bio fuel cell, all you would have to do is add sugar and water.”


Human  Heart



Summary prepared by Michael T. Allen, University of Southern Mississippi – Gulf Coast, in collaboration with the Psychosocial Working Group.

Definition and Background

The cardiovascular system functions to provide nutrients to systemic tissue beds of the body, as well as to remove waste products of cellular metabolism. In order to accomplish this formidable responsibility, the heart and vasculature must work in concert and be flexible enough to respond to a wide range of activities, ranging from quiet rest or sleep to maximal exercise. Thus, the cardiovascular system is continuously “reactive,” depending on the metabolic needs of the organism.

The use of the term “cardiovascular reactivity” by researchers in the field of cardiovascular behavioral medicine or psychophysiology generally is defined more narrowly than that described in the previous paragraph. That is, cardiovascular reactivity is usually understood to reflect the physiologic changes from a resting or baseline state to some type of psychological or physical challenge or stressor. Importantly, it is widely thought that individuals showing exaggerated cardiovascular responses to these stressful conditions may be more at risk for the development of cardiovascular syndromes such as hypertension or coronary heart disease than those exhibiting relatively smaller responses. It should be pointed out that how one defines “exaggerated responses” has ranged from simply taking individuals who show the largest responses in their study group, to more rigorous attempts to define reactors in terms on whether the cardiovascular responses exceed the metabolic demands of the situation. Regardless, the underlying assumption is that large increases in cardiovascular responses to stressors that occur frequently may lead to alterations in either the heart or vasculature that can have deleterious effects on the individual’s health.

A number of epidemiological studies have pointed to an inverse relationship between SES and health outcomes; that is, lower SES is associated with increased risk for a number of diseases. This relationship has also been found for SES and cardiovascular disease. Although multiple factors such as diet, compliance and access to health care have been postulated, a popular conceptualization of the mechanisms linking SES and health outcomes is based upon the observations that individuals from low SES environments generally experience more day-to-day stress than individuals living in more affluent locales. This differential stress exposure along the SES gradient has implications for how one might view reactivity as either a moderator or mediator of the relationship between SES and health outcomes.

If cardiovascular reactivity during stress is a consistent physiological characteristic of an individual, then one might expect that highly reactive individuals who live in high stress environments would have a greater stress load (more frequent and greater physiological stress responses) than individuals who are either not highly reactive or live in lower stress areas. In this case, one could conceptualize reactivity as interacting with or moderating the effects of environmental stress as a link between SES and health outcomes. Another possibility for the role of reactivity as a link between SES and health outcomes is that exposure to a more threatening or challenging environment by lower SES individuals results in greater reactivity in various organ systems in response to this exposure. Over time, individuals who are chronically exposed to more threatening environments may also begin to anticipate or expect threats even in benign situations, leading to greater overall stress load on manifold organ systems. Here, reactivity plays more of a mediating role between SES/stress exposure and health outcomes. These two possibilities are not mutually exclusive; both could operate to some extent in different individuals. This summary will address both of these possibilities in later sections. Let us first turn our attention to the measurement of cardiovascular reactivity.


How one chooses to measure cardiovascular reactivity is a complex issue due to a number of potential cardiovascular variables, as well as a variety of ways in which to measure change.

Among the cardiovascular measures that have been utilized in reactivity studies are heart rate, blood pressure (systolic, diastolic, mean arterial), stroke volume (average amount of blood pumped from the left ventricle on a given contraction of the heart), cardiac output (volume of blood pumped per minute), total peripheral resistance (resistance to blood flow by the systemic vasculature – a derived measure computed when cardiac output and mean arterial pressure are known), and timing intervals in the cardiac cycle that reflect cardiac performance such as pre-ejection period (PEP) and left ventricular ejection time (LVET). Over the years, the journal Psychophysiology has published methodological guidelines for measuring heart rate, blood pressure, and impedance cardiography (a noninvasive technique for measuring stroke volume, cardiac output, total peripheral resistance, and systolic time intervals). The reader is referred to these papers for detailed discussions of measurement procedures for these cardiovascular variables.

Although not always explicitly stated, many reactivity studies are interested in how the autonomic nervous system responds to environmental challenges, and the resultant effects on the cardiovascular system. For example, heart rate reflects both sympathetic and parasympathetic (vagal) influences on the sino-atrial node, whereas PEP is most directly influenced by sympathetic influences on contractility of the ventricles of the heart. Systolic blood pressure reflects both increases in contractility of the ventricles and the amount of systemic resistance to blood flow, whereas diastolic pressure is more reflective of vascular resistance (as is of course total peripheral resistance). The variability of heart rate has also become a popular measure, as the variability of beat-to-beat heart periods may reflect the degree of vagal control of the heart under certain conditions. A number of computational procedures to measure heart rate variability have been utilized such as spectral analysis and sequential differencing techniques.

Thus, studies of cardiovascular reactivity to stress not only allow for examination of cardiovascular dynamics, but also give a window for assessing autonomic nervous system adjustments to these environmental demands.

Although the measurement of reactivity as reflecting the degree of change from a baseline period to some period of challenge seems simple enough, there is disagreement concerning the best way to measure “change.” This disagreement often stems from how to interpret the amount of change exhibited by groups when their base rates are different. Some have championed the use of simple or raw change, whereas others have called for the use of “residualized” change scores which are computed from regression analyses of baseline and task levels. A detailed discussion of these issues is beyond the scope of this summary; the reader is referred to Wainer (1991) and Maxwell et al  for interesting observations on these issues.

It has become increasingly clear to researchers that the use of only one or two cardiovascular measures often gives an incomplete picture of autonomic adjustments that occur during laboratory challenges or everyday life. An emerging strategy is to measure a number of cardiovascular variables to examine the pattern of responses. That is, rather than look at the response of a number of variables independently, it is recognized that the cardiovascular system responds with a limited number of organized patterns of response. Remember that the cardiovascular system has a very specific and important job: to get the appropriate amount of blood flow to various vascular beds according their needs. The various measures of cardiovascular function would not be expected to vary randomly among each other, but be organized in a finite number of patterns that respond to various environmental demands. Individuals are therefore studied based upon their composite pattern of autonomic and cardiovascular response, rather than study individuals on the basis of their reactivity to cardiovascular variables in isolation. This strategy promises to provide more useful information on neural control of cardiovascular response during both rest and stress.

One of the fundamental assumptions of the importance of stress-related cardiovascular reactivity and cardiovascular disease is that reactivity exhibited by individuals shows consistency over time. This has traditionally been referred to as individual response stereotypy. Most theories of the manner in which reactivity can affect health outcomes assume that repeated exaggerated cardiovascular responses to stress may trigger maladaptive physiological processes. This implies a consistency of response when confronted by similar challenges at different points in time. Thus, studies have examined the temporal stability of cardiovascular response to laboratory or “real-life” challenges. Although results have varied, most studies have found good consistency across time periods for measures such as heart rate, systolic blood pressure, and impedance cardiography-derived variables.

Another reliability issue in the measurement of cardiovascular reactivity has been the degree to which a given challenge will produce a similar pattern of response in different individuals, i.e., situational response stereotypy. For instance, will a given laboratory stressor produce a similar pattern of cardiovascular response in different people? Studies examining this question have shown that there generally are a number of patterns of response exhibited by different individuals, but there is a modal response that most individuals exhibit during a stressor. Thus, there are certain tasks that are more likely to elicit a particular pattern of autonomic and cardiovascular response than others. In an effort to help researchers select tasks that are more likely to produce a given pattern of response.


The issue of selection of an appropriate task to elicit cardiovascular reactivity has special relevance when considering the role of reactivity as a potential mediator between SES and health. As mentioned earlier, individuals living in low SES environments generally are exposed to more day-to-day stress than high SES individuals. Interestingly, one could speculate that greater chronic stress exposure could have either an accentuating or attenuating influence on acute stress exposure in cardiovascular reactivity paradigms. In the first instance, chronic exposure to stress would be conceptualized as already taxing a person’s ability to cope with new, acute challenges. Individuals experiencing high levels of background chronic stress would show exaggerated acute stress responses (greater cardiovascular reactivity in this case) as compared to individuals exposed to minimal chronic stress. On the other hand, one could also envision that response to chronic stress might have a dampening effect on acute stress responses. This could be due to individuals having the time and opportunity to learn adaptations to stress.

A recent review of studies addressing the effects of background stress on acute reactivity and recovery from stress in a total of 19 studies indicated equivocal results, although heightened acute reactivity was found in a slight majority of the studies. Among the many problems with the interpretations of these studies is the accurate measurement of background stress. Individuals within each study also show a wide range of individual differences in their acute responses, regardless of background stress. Clearly this is an unresolved issue that will require much additional study.

Finally, the “ecological validity” of measures of lab-based reactivity has been questioned. That is, some have argued that reactivity observed in sometimes contrived laboratory situations might not generalize to the “real world” and to the types of situations to which a person may be exposed in everyday life. This is especially important as it relates to differences in SES. Many of the commonly used laboratory tasks such as mental arithmetic or other problem-solving tasks may be differentially challenging to research participants who vary in intelligence and/or academic achievement. A person who is moderately challenged by the task may show more reactivity than one who is only minimally challenged. On the other hand, a person who finds the task to be exceedingly difficult may disengage from the task and show very little reactivity. Individuals from low SES environments may show more task disengagement because of a lack of adequate academic preparation, or they could become more easily frustrated because of the background chronic stress to which they are exposed. It is important to develop these tasks with built-in adjustments for task delivery so that the difficulty of the task will be roughly equivalent for all individuals. Tasks that require the individual to talk about unpleasant or stressful events in their lives can also vary considerably depending on SES. A high SES individual may discuss a situation concerning a rather benign disagreement at home or at school, whereas an individual living in a high crime area may discuss a shooting outside of his/her home. The point is that the researcher needs to consider whether the chosen task(s) to elicit acute reactivity may be differentially interpreted depending on SES status and differences in chronic stress.

Even if steps are taken to choose or modify tasks so as to minimize potential SES differences, there is still the issue of how well the reactivity seen during laboratory stressors index the reactivity exhibited by individuals in real-life interactions. Accordingly, a number of studies have compared laboratory-based reactivity with ambulatory responses, most often ambulatory blood pressure. The ambulatory monitors used in these studies are devices that usually measure both blood pressure and heart rate, but are small enough to be worn unobtrusively by the individual. These studies are important in trying to establish that laboratory-based reactivity is representative of the magnitude of response that individuals exhibit in everyday interactions.

Although there are conceptual difficulties in trying to find equivalent ambulatory periods with which to compare laboratory stress or resting periods, many studies have found acceptable levels of correspondence between laboratory and ambulatory levels. For example,Lindenand Con  reported that an overall average of SBP reactivity during three laboratory challenges was a significant predictor of ambulatory blood pressure mean. Other studies have tried to specify more precisely the ambulatory periods that were likely to correspond to lab-based reactivity. Matthews et al.  reported that the correspondence between ambulatory and lab BP values was strongest during the ambulatory periods of perceived stress. Thus, their conclusion was that ambulatory and lab-based BP responses were related, but one needed to take into account that the relationship is strengthened when appropriate ambulatory periods (such as times when the person is experiencing stress) are chosen. Steptoe et al.  echoed these findings by reporting that ambulatory/lab associations were more consistent when the level of perceived stress and physical activity in the lab and field situations were more congruent. These studies point out that investigation in this area must go beyond merely correlating lab measures with overall ambulatory responses. To address whether lab-based reactivity can be generalized to the “real world,” it is prudent to pick ambulatory periods that are similar in perceived stress to that experienced during the laboratory challenges.

Relation to SES

What evidence is available to help one understand the potential relationship between cardiovascular reactivity and SES? To date, there have not been a large number of studies that have explored this issue, although a few informative studies are available. Lynch, Everson, Kaplan, Salonen and Salonen examined whether low SES and heightened cardiovascular reactivity had interactive effects on the progression of carotid atherosclerosis in men enrolled in theKuopio,Finlandstudy. In this study, cardiovascular reactivity was defined as the increase in SBP response in anticipation of a maximal exercise stress test. Results indicated that the greatest progression of atherosclerosis occurred in men who had both heightened reactivity and low SES. Although this study did not examine directly the relationship of SES and reactivity, the results with atherosclerosis risk suggest a natural clustering of SES and reactivity in determining a negative health outcome, rather than being independent factors. It is also of interest that an interaction was found; that is, the greatest progression of atherosclerosis was in a group that was not only low SES, but who were also more highly reactive. This suggests a moderating influence of reactivity on the relationship between SES and health outcomes.

A recent study by Gump et al.  used structural equation modeling to examine the relationships among SES and cardiovascular reactivity in Black and White children. SES was defined in two ways: family SES was measured using the Four Factor Index of Social Status as devised by Hollingshead, and neighborhood SES was determined using information from census tract data such as educational attainment, percent of single mothers, and population density. The models relating SES and reactivity were different for Blacks and Whites. For Blacks, both neighborhood and family SES were negatively related to reactivity (higher SES associated with lower reactivity), with the relationships being mediated by hostility as measured by the Cook-Medley Ho scale. For Whites, family SES was negatively related to reactivity, although neighborhood SES was not. Interestingly, the family SES/reactivity relationship was not mediated via hostility. The reader is referred to Gump et al. for a detailed discussion of these intriguing findings. Although this study does not relate either SES or reactivity to a health outcome, the study does suggest that the relationship of SES and reactivity may also be modulated by other influences such as ethnicity or hostility.

The finding of increased reactivity being related to lower SES has not been found in all studies. Data from the Whitehall II study indicated that SBP increase during a mental stress task was associated with higher occupational grade. Gump et al.  have speculated that this finding in the Whitehall II study may have been due to the use of the Raven’s Progressive Matrices, a nonverbal intelligence test, as the mental stress. They suggest that this may have produced more effort and challenge in the high occupational group, and disengagement in the lower occupational group. The need to make sure that the laboratory challenges are as equivalent as possible for individuals from different SES levels was discussed in the last section.

Another line of research may indirectly point to a possible relationship between SES and reactivity. This is the study of race differences between Blacks and Whites with regard to cardiovascular reactivity. Studies in this area have usually either explicitly or implicitly assumed that any racial differences in reactivity were due to genetic differences, and in fact some differences in the baseline levels of heart rate and blood pressure have been found between Black and White infants. Yet, it has been persuasively argued that social environment is a much stronger factor for racial differences than genetics. That is, some have suggested that race may be more accurately seen as a proxy for differences in SES between Blacks and Whites. Viewed in this manner, racial differences in reactivity may shed light on SES and reactivity.

To date, a number of studies have found greater vascular reactivity responses in Blacks than Whites. There is evidence that Blacks may have more alpha-adrenergic responses to stress than Whites, with Blacks also exhibiting a blunted beta-adrenergic response. These stronger vascular responses are consistent with observations concerning the natural history of essential hypertension and the fact that Blacks have a higher incidence of hypertension and coronary heart disease than age-matched White counterparts. It is also interesting that a study by Allen and Matthews examined the effects of race on cardiovascular reactivity in a sample of children and adolescents in which the authors attempted to match Black and White participants on SES. Although the matching was not completely successful in equating SES in the groups, SES differences were less than in most reactivity studies examining race differences in young people. Interesting, no race differences in vasoconstrictive responses were found in this study. This is consistent with the notion that at least some of the racial differences in vasoconstrictive responses reported in other studies may be related to lower SES in the Black samples. One might also speculate that reactivity differences are not as easily found in samples of children and adolescents as in adult samples due to the longer exposure to the stress of lower SES lifestyles and neighborhoods in adults. Obviously, much additional research is needed to further illuminate these issues.

Relation to Health

Although the research question of the importance of exaggerated cardiovascular responses to stress for deleterious cardiovascular health outcomes has been asked for many years, the direct evidence for this relationship is still sparse. One of the obvious reasons for this is that longitudinal studies are needed to effectively address these questions. Individuals with already established cardiovascular disease cannot be used as the disease process itself will alter the cardiovascular responses to stress. The area that has been most commonly addressed is the degree to which exaggerated cardiovascular reactivity is a risk factor for hypertension.

Most of the longitudinal studies to date have used reactivity to dynamic exercise or the cold pressor test as predictors for hypertension. Studies such as Jackson et al.  have reported that future hypertension in previously normotensive adults was predicted by blood pressure responses to dynamic exercise. The cardiovascular responses of individuals during the cold pressor test (immersion of hand or foot in ice water) have been examined longitudinally in a few studies. For example, Menkes et al. report the findings on 910 White, male medical students who had blood pressure and pulse rate measured before and during a cold pressor test. A significant association was found between maximal change in SBP during the cold pressor test and subsequent development of hypertension after many years. This relationship persisted even after controlling for such factors as cigarette smoking, initial resting SBP, and family history of hypertension. Data from the Bogalusa Heart Study on children indicate that peak cardiovascular reactivity during three physical challenges (orthostatic challenge, handgrip exercise, and the cold pressor) predicted SBP and DBP resting levels four years later. It should be noted that some studies have not found this relationship. For example, Carroll et al.  report that DBP reaction to the cold pressor test in a group of 1039 men explained only a very small portion of the variance in follow-up DBP after 5 years (SBP had no explanatory power). However, the men in the Carroll et al. study averaged 56.6 years of age, whereas the participants in the Menkes et al. study were young medical students. As most researchers posit that excessive reactivity may predict neurogenic hypertension that may have a relatively early onset, the young sample in the Menkes study may be a more appropriate sample.

The various hypotheses concerning the potential role of psychological stress as a factor in the etiology of cardiovascular disease have generally emphasized the deleterious effects of excessive sympathetic nervous system activation on the cardiovascular system, although parasympathetic withdrawal has also been considered. One stimulus for this line of inquiry has been animal studies reporting sustained elevations in blood pressure in certain genetic strains of rats who were subjected to stressful conflict avoidance situations, as well as elevated blood pressure in mice who were housed in overcrowded conditions. Therefore, a few longitudinal studies have examined the predictive power of responses to psychologically challenging situations on later blood pressure status.

Borghi et al.  reported that the magnitude of the DBP responses during a mental arithmetic challenge significantly predicted the increase in blood pressure after 5 years in young borderline hypertensive individuals. Everson et al, in a study of middle-aged men in Kuopio, Finland, investigated whether the rise in BP in anticipation of a bicycle ergometer stress test was related to subsequent longitudinal increases in resting BP. Results indicated that men exhibiting either an anticipatory SBP response greater than or equal to 30 mm Hg or a DBP response greater than or equal to 15 mm Hg had nearly four times the risk of developing hypertension after a 4 year period. This relationship was maintained even after controlling for traditional risk factors for hypertension. Light et al. reported that SBP and heart rate responses to a challenging reaction time task were significant predictors of blood pressure status after a 10-15 year period. The participants in this study were college students at the initial testing, and all initially were normotensive males. In analyzing data from the CARDIA study on young normotensive men and women, Markovitz et al. report that increased SBP reactivity during a challenging video game was associated with increased SBP after 5 years, independent of resting SBP. Interestingly, this relationship was found to hold for men but not women. This study did not find a predictive relationship for reactivity during the cold pressor or mirror tracing task and subsequent BP levels.

Matthews et al.  examined the prognostic value of cardiovascular reactivity for follow-up blood pressure in male and female children. The measure of reactivity included a composite standardized reactivity score during mirror tracing, mental arithmetic and isometric handgrip. Among boys only, larger blood pressure reactivity scores were associated with greater blood pressure levels after 6 ½ years. Similar findings using a video game as the psychological stressor were found by Murphy et al. . Thus, there are a number of studies that have reported that some aspect of cardiovascular reactivity does predict later levels of blood pressure, although the relationship may be stronger for males than females.

Although most studies in this area have examined the relationship of cardiovascular reactivity to physical or psychological challenges and future increases in BP or development of hypertension, a few studies have investigated whether reactivity is related to cardiovascular disease progression in individuals who already have documented disease. For example, Manuck et al. report the results of a pilot study on 14 individuals who had already experienced a myocardial infarction. Five of these individuals subsequently suffered either a reinfarction or stroke. These five individuals had previously shown significantly larger increases in SBP and DBP during a frustrating Stroop Color-Word Interference task than the remaining individuals who had not experienced a second clinical event. In a related study, Barnett et al.  report that the amount of carotid artery plaque development over a 2 year period in a group of 136 untreated patients was significantly related to SBP reactivity during the Stroop test, along with a number of traditional risk factors. These studies suggest that cardiovascular reactivity to psychological stress can influence the development of atherosclerosis in susceptible patients.






A number of epidemiological studies have found that lower SES is associated with increased risk for a negative cardiovascular health outcome. To what extent can reactivity protocols help one to understand the SES-health gradient? Does reactivity serve a moderating role whereby individuals who are both highly reactive and experience the burden of a chronic stressful environment are at greater risk for deleterious health outcomes? At least one study from theKuopio,Finlandstudies points to this moderating role of reactivity. There is also the possibility that chronic exposure to stress by low SES individuals leads to a developed pattern of responding to the environment with anticipated threat or challenge. The concurrent heightened physiological reactivity during these behavioral patterns may, over time, result in negative health outcomes. This is consistent with findings from studies like Gump et al. which finds that heightened vasoconstrictive responses are associated with lower SES. But it is also the case that ethnicity and/or characteristics like hostility may moderate these relationships. Thus, reactivity may link SES and health outcomes, although the precise role of reactivity will require much more study. This role may be in conjunction with, rather than in lieu of, other potential mediators of the SES/health connection such as diet, availability of medical care, adherence to treatment regimens, etc.

There are a number of factors that make it difficult to test the role of reactivity in linking SES and health outcomes. First, the definition of reactivity has varied from study to study. Whether to use heart rate, blood pressure, cardiac performance variables, vascular resistance, or a combination of multiple measures into cardiovascular patterns, has varied from study to study. Other factors are whether to use laboratory-based or ambulatory-based challenges. How to measure change from a baseline period to a task or challenge period has also been debated. One thing does seem clear. For the field of reactivity research to progress, researchers, whether in psychology, behavioral medicine, health psychology or others, examining reactivity relationships with SES and/or health outcomes must understand basic cardiovascular physiology rather than conceive of cardiovascular variables such as heart rate and blood pressure as ends in themselves. It will not be enough to simply report an increase in, say, blood pressure; rather, one must try to understand the adjustments made by the entire cardiovascular system and the autonomic underpinnings of those adjustments. We must, as the late Paul Obrist was oft to say, become “better biologists”.

Finally, the total reliance on large-n studies to elucidate the patterns of cardiovascular adjustments to challenge is being challenged on a number of fronts. The traditional way of comparing large groups of individuals to discover characteristic cardiovascular patterns of response results in data that have been “homogenized” across a large number of individuals. These mean responses of groups give one a stable, average characterization of response patterns, but much of the richness of the data from individuals is lost. A growing number of researchers  are calling for a more “idiographic” approach to understanding cardiovascular dynamics during stress. The strategy here is to study a relatively small number of individuals, but to do extensive evaluation of these individuals under a number of challenging and baseline conditions. This “case study” approach, as opposed to the more “epidemiological” approach of the large-n studies, may more accurately assess the pattern of cardiovascular adjustments exhibited by individuals. This is not to say that the traditional large-n studies should be replaced by the idiographic studies; the development of statistics for the small-n studies has not progressed as rapidly as the traditional designs, and most funding sources such as the National Institutes of Health are not accustomed to funding studies where power determination is unclear. Rather, the small-n studies would complement the findings from the large-n studies. This is an important conceptual area to be addressed when looking at the future of cardiovascular reactivity research.




Endothelial Function Testing as a Biomarker of Vascular Disease


Endothelium is the lining of blood vessels




  1. 1.   Subodh Verma, MD, PhD;
  2. 2.   Michael R. Buchanan, PhD;
  3. 3.   Todd J. Anderson, MD

+ Author Affiliations

  1. 1.   From the Department of Cardiac Surgery, University of Toronto, Toronto, Ontario, and Department of Pharmacology and Therapeutics, University of Calgary, Calgary, Alberta, Canada (S.V.); the Department of Pathology and Molecular Science, McMaster University, Health Sciences Centre, Hamilton, Ontario, Canada (M.B.); and the Department of Medicine, University of Calgary, Calgary, Alberta, Canada (T.J.A.).
  2. Correspondence to Subodh Verma, MD, PhD, Division of Cardiac Surgery, 14EN-215, 200 Elizabeth St, Toronto General Hospital, Toronto, Ontario, Canada M5G 2C4. E-mail



The endothelium is the monolayer of endothelial cells lining the lumen of all blood vessels. These cells function as a protective biocompatible barrier between all tissues and the circulating blood. Endothelial cells also function as a selective sieve to facilitate bidirectional passage of macromolecules and blood gases to and from tissues and blood. The strategic location of the endothelium allows it to “sense” changes in hemodynamic forces and blood-borne signals and “respond” by releasing a number of autocrine and paracrine substances. A balanced release of these bioactive factors facilitates vascular homeostasis. Endothelial cell dysfunction disrupts this balance, thereby predisposing the vessel wall to vasoconstriction, leukocyte adherence, platelet activation, mitogenesis, pro-oxidation, thrombosis, impaired coagulation, vascular inflammation, and atherosclerosis.1 Our understanding of these endothelial cell responses has led to the development of tests that are believed to reflect endothelial cell dysfunction or integrity in vivo. Given the central role of the endothelium in the development and clinical course of atherosclerosis, endothelial function testing may serve as a useful biomarker of atherosclerosis.

Nitric oxide (NO) is the key endothelium-derived relaxing factor that plays a pivotal role in the maintenance of vascular tone and reactivity.2 In addition to being the main determinant of basal vascular smooth muscle tone, NO acts to negate the actions of potent endothelium-derived contracting factors such as angiotensin II and endothelin-1. In addition, NO serves to inhibit platelet and white cell activation and to maintain the vascular smooth muscle in a nonproliferative state. NO is synthesized from l-arginine under the influence of the enzyme NO synthase (NOS). NOS requires a critical cofactor, tetrahydrobiopterin, to facilitate NO production. Tetrahydrobiopterin deficiency leads to an “uncoupling” of NOS that results in the formation of untoward oxidants such as superoxide and hydrogen peroxide (versus NO) with resultant impairment in endothelial function.3 Superoxide inactivates NO to peroxynitrite, which further decreases NO activity in this uncoupled state. Cardiac risk factors in general lead to an increase in oxidative stress, attenuating net NO bioactivity.

Assessment of Endothelial Function

Assessment of endothelial cell function refers to a measure of endothelial cell response to stimulation—for example, by vasoactive substances released by or those that interact with the vascular endothelium. Endothelium-dependent vasodilation can be assessed in the coronary and peripheral circulations in humans. In addition, measures of platelet function and inflammation/leukocyte activation (such as C-reactive protein [CRP]) are other measures of endothelial health. Figure 1 shows methods of assessing endothelium-dependent vasomotion.


Figure 1. Assessment of endothelium-dependent vasomotion can be achieved both invasively and noninvasively in coronary and peripheral circulations. In patients with healthy endothelial function, endothelium-dependent stimuli (eg, flow or Ach) incite vasodilatation primarily by augmenting the release of NO (normal endothelial vasomotion). In patients with risk factors for coronary disease or with established atherosclerosis, endothelial dysfunction becomes apparent by decreased or paradoxical responses to endothelium-dependent stimuli. Emerging data support the concept that assessment of endothelial vasomotion may be a useful biomarker for atherosclerotic vascular disease. A-II indicates angiotensin II; ET-1, endothelin-1; and FMD, flow-mediated dilatation.

Coronary Circulation

Quantitative coronary angiography has been used to examine the changes in vascular diameter in response to an infusion of an endothelium-dependent vasodilator such as acetylcholine (Ach), bradykinin, substance P, or serotonin.4 In healthy vessels, Ach evokes a NO-mediated vasodilatory response; however, in patients with endothelial dysfunction, this effect is blunted or paradoxical vasoconstriction may occur.5 Endothelial function of the coronary microvasculature (resistance vessels) has been assessed by using intracoronary Doppler techniques to measure coronary blood flow in response to pharmacological or physiological stimuli.6 Although considered by many to be the best assessment of endothelial function, this technique is limited by its invasive nature, expense, and relative inaccessibility.

Peripheral Circulation

Brachial artery ultrasound is a widely used, noninvasive measure of endothelial cell function.7,8 The forearm blood flow is occluded for 5 minutes using a blood pressure cuff maintained at a standard pressure. When the pressure is released, reactive hyperemia occurs. This results in shear stress-induced NO release and subsequent vasodilatation (flow-mediated vasodilatation). This technique has the advantage of being noninvasive and can readily identify populations with attenuated endothelial function. The major limitations of this technique are the need for ultrasonographic expertise and a significant day-to-day variability (about 25%) due to biological circadian rhythms. Nonetheless, at present, this approach is widely used to assess vasomotion function.

Resistance vessel function in the forearm is assessed by strain-gauge venous impedance plethysmography. This methodology examines the change in forearm blood flow in response to direct intraarterial (brachial artery) administration of agonists. This technique is excellent for acute interventions with repeated measurements.9,10 The major drawbacks, again, are reproducibility and the technique’s more invasive nature compared with ultrasound. Noninvasive measures of arterial compliance and waveform morphology also provide a marker of vascular health that may in part be endothelium dependent.

Inflammatory Markers

Over the past few years, we have witnessed a paradigm shift in our understanding of the underlying principles of atherosclerosis. This new view supports the concept that vascular inflammation is a central orchestrator of atherosclerotic lesion formation, progression, and eventual rupture.11 Chronic inflammation results in endothelial dysfunction and facilitates the interactions among modified lipoproteins, monocyte-derived macrophages, T cells, and normal cellular elements of the arterial wall, thus inciting early and late atherosclerotic processes. This paradigm has fueled interest in evaluating inflammatory markers of atherosclerosis, of which high-sensitivity CRP has emerged as one of the most important. As such, CRP is a powerful independent predictor of myocardial infarction, stroke, and vascular death in a variety of settings and appears to be a better prognosticator of cardiovascular events than LDL cholesterol.12–14 Over the past year, much interest has been generated into unraveling the mechanistic basis of the CRP-atherosclerosis connection (Figure 2). Indeed, recent studies, including work from our laboratory, suggest that CRP is not only a predictor but also a mediator of lesion formation.15–25 CRP, at concentrations known to predict vascular disease, has a direct effect to stimulate diverse early atherosclerotic processes, including the expression of endothelial cell adhesion molecules, the production of chemoattractant chemokines, and macrophage LDL uptake. In addition, CRP directly modulates the production of endothelium-derived vasoactive factors, including downregulating endothelial NOS (eNOS)-derived NO while augmenting production of the potent endothelium-derived vasoconstrictor endothelin-1. Additionally, CRP facilitates endothelial cell apoptosis and attenuates angiogenesis, which is an important compensatory mechanism in ischemia. More recently, CRP has also been demonstrated to promote the release of plasminogen activator inhibitor-1 from endothelial cells,25 upregulate angiotensin-mediated neointimal formation,20 and alter endothelial progenitor cell survival and differentiation.22


Figure 2. The mechanistic basis of the predictive value of CRP may be its ability to incite endothelial dysfunction. CRP can decrease eNOS mRNA,16 augment endothelin-1 (ET-1),15 and upregulate diverse adhesion molecules and chemoattractant chemokines, uncovering a proinflammatory and proatherosclerotic phenotype. Preliminary observations also suggest that CRP upregulates the nuclear factor-κB signaling in endothelial cells while attenuating endothelial progenitor cell survival and differentiation (Verma et al, unpublished observations, 2003). Also, the proatherogenic effects of CRP are augmented in the hyperglycemic milieu.21,25 Also, CRP has been demonstrated to potently upregulate angiotensin type 1 (ATI-R) receptor in vascular smooth muscle cells in vivo and in vitro, augmenting vascular smooth muscle (VSM) proliferation and migration, reactive oxygen species (ROS) production, and restenosis.20 CRP therefore seems to function as an important circulating marker of endothelial dysfunction. ICAM indicates intercellular adhesion molecule; VCAM, vascular cell adhesion molecule; and MCP, monocyte chemoattractant protein.



Thus, CRP is not only an inflammatory marker of atherosclerosis/coronary events but is also a mediator of the disease because it contributes to the pathogenesis of lesion formation, plaque rupture, and coronary thrombosis by interacting with and altering the endothelial cell phenotype. CRP should be regarded as an indirect, albeit important, measure of endothelial function. The major advantage of biochemical measures of endothelial function is that they are inexpensive and offer excellent reproducibility. However, it is not yet clear what the relationship is between these markers, measures of endothelium-dependent vasodilation, and cardiovascular outcomes, and it is an area that is currently being investigated. Preliminary studies suggest that patients with elevated levels of CRP have impaired endothelium-dependent vasodilatation, which suggests that CRP may be a useful clinical tool for endothelial vasomotion.26

Endothelial Function as a Biomarker of Atherosclerosis

Dysfunction of endothelial cells is probably the earliest event in the process of lesion formation—hence, the concept that assessment of endothelial function may be a useful prognostic tool for coronary artery disease. Heterogeneity of vascular dysfunction must be appreciated. Nonetheless, coronary endothelial cell perturbations often are reflected in peripheral vasodilator abnormalities, thereby allowing the assessment of peripheral endothelial function as a measure of coronary vasomotion.27 Recent studies also suggest that there is a correlation between endothelium-dependent vasodilation and CRP levels.26 Thus, endothelial dysfunction may be reflected systemically, thereby allowing for a less invasive approach to the assessment of overall endothelial cell biocompatibility.

The rationale for why endothelial function testing, either directly or indirectly, may serve as an indicator of vascular damage or disease is as follows: (1) The healthy endothelium is nonthrombogenic; (2) endothelial dysfunction occurs in response to vascular risk factors and is an early event in atherosclerosis; (3) endothelial dysfunction precedes structural atherosclerosis; (4) interventions that improve endothelial function also decrease cardiovascular events in patients with stable coronary disease; (5) reproducible, noninvasive assessments of endothelial function exist; and (6) endothelial function testing fulfills the criteria for an acceptable biomarker. Because risk factors, even those not yet identified, target the endothelium, it is a logical “window” of future atherosclerotic outcomes.

Prognostic Relevance of Endothelial Function

The clinical manifestations of coronary artery disease depend on a multitude of interrelated pathophysiological processes, of which endothelial dysfunction is only one.

Studies of Vasomotion

We demonstrated that endothelial dysfunction was associated with the development of transplantation atherosclerosis as assessed by intravascular ultrasound.28 Those subjects with normal vasodilator responses to Ach immediately after transplantation developed atherosclerosis at a rate one third that of those with endothelial dysfunction during the first year of follow-up. Three retrospective trials have assessed the relationship between Ach-mediated coronary endothelial function and clinical events (Table). Subjects with vasoconstrictor responses to Ach were more likely to develop adverse cardiovascular events during follow-up of 5 to 10 years despite having minimal coronary disease at baseline. Although important, these observations are diminished by the relative low frequency of end point events.29,30 A recently published trial of 308 subjects who underwent coronary endothelial function testing revealed improved survival in those subjects with better conduit and resistance vessel responses to Ach. Endothelial function was a predictor of outcome in a multivariate analysis.31

Studies of Prognostic Implications of Endothelium-Dependent Vasomotion

Perticone and colleagues32 studied 225 hypertensive subjects who underwent Ach testing in the forearm with plethysmography. After correcting for blood pressure, subjects with the lowest tertile of endothelial function had an increase in cardiovascular events over a 3-year follow-up. The most robust of the prognostic studies was reported by Heitzer et al,33 who studied 281 subjects with coronary disease undergoing forearm Ach studies. Two key observations were made. First, subjects with attenuated responses had more events. Second, subjects with a greater acute improvement of endothelial function with vitamin C (suggesting more oxidative stress) had a worse outcome.

Small studies have also suggested a prognostic role for the brachial ultrasound assessment of endothelial function.34 A recently reported prospective study demonstrated that abnormalities of flow-mediated dilation were predictive of postoperative complications in patients undergoing noncardiac vascular surgery.35

All the studies done to date are too small to be definitive. Studies with thousands of subjects are underway36 with brachial ultrasound to determine if a single measure of vasoreactivity in an individual patient predicts the development of atherosclerosis or its complications (J. Vita, MD, personal communication, 2002).

Studies of Inflammatory Markers

Studies by Ridker37 have demonstrated a relationship of cardiovascular outcomes with interleukin-6, tumor necrosis factor-α, soluble P selectin, and soluble intercellular adhesion molecule-1.38 However, high-sensitivity CRP has emerged as the most powerful predictor of future cardiovascular events.12–14

In both healthy men and women, CRP levels in the upper quartile increase risk of adverse clinical events by 2- to 4-fold and are at least as prognostic as lipid parameters. Of great interest is the fact that subjects with elevated levels of CRP seem to gain more benefit from pharmacological therapies such as aspirin or statin therapy. Given the direct effects of CRP to destabilize eNOS mRNA in endothelial cells,16 it is logical to propose that CRP is a sensitive marker of endothelial dysfunction; it remains to be determined whether the predictive value of endothelial vasomotion assessments will be independent of CRP levels.

An alternative approach to measuring endothelial cell function in response to assorted agonists is to measure changes in endothelial cell biocompatibility. In support of this possibility, Brister et al38 found that the lipoxygenase pathway-derived monohydroxide, 13-hydroxyoctadecadienoic acid (13-HODE) decreases with age in patients with coronary artery disease undergoing coronary artery bypass grafting and that this decrease is associated with increased endothelial cell thrombogenicity.38 More recently, 13-HODE plasma levels have been shown to dramatically increase in patients with confirmed atherosclerosis and were higher in coronary artery bypass grafting patients who suffered a cardiovascular event during the 2-year follow-up.39

Several other surrogate markers of endothelial cell activation and inflammation are being investigated, including lipoprotein-associated phospholipase A2,40 CD40 receptor/CD40-ligand interaction, 41 LOX-1,42 and measurement of circulating endothelial progenitor cells.43,44

Implications for Practice

Atherosclerotic vascular disease remains the leading cause of morbidity and mortality among adults in developed countries and is increasing at an alarming rate in developing nations. The concept of risk introduced by the Framingham Heart Study more than 50 years ago serves as the “gold standard” in risk assessment.45 Enthusiasm for risk assessment46 and prevention are based on the demonstration that aggressive medical therapy reduces the likelihood of recurrent coronary events in patients with established coronary heart disease. A similar potential exists for primary prevention.

Patients at high risk (those with diabetes, other vascular disease, or multiple risk factors) clearly benefit from pharmacological therapy. In addition, patients without traditional risk factors are at low risk of events over both the short and long term. It is estimated however, that 40% to 50% of adults fall into the intermediate risk group according to National Health and Examination Survey (NHANES) III data. Patients in this intermediate risk group do not currently qualify for the most intensive risk factor interventions. The problem with this approach is that the treatment algorithms take into account short-term (10 years) rather than long-term risk (30 years), and there is a wide range of risk within this large group. It is this group that could potentially benefit from further risk stratification with endothelial function testing. If prospective studies confirm the predictive nature of endothelial markers for cardiovascular outcomes, then incorporation of these measures into risk factor models would lead to more effective prevention. A positive test in a subject at low to moderate risk would identify an individual whose risk would warrant pharmacological treatment.47 Other proposed markers of risk, including coronary calcium score (electron beam computed tomography), carotid intimal-medial thickness, ankle-brachial index, and stress testing, are also being evaluated carefully. The optimal diagnostic/therapeutic approach is not known at this time but will be better defined in the next decade. These data will come from ongoing prospective studies of endothelial function and the National Institutes of Health-sponsored MESA study (Multi-Ethnic Study of Atherosclerosis). Given the increasing atherosclerosis pandemic, advances in diagnostic modalities, and health economics issues, the time is right to define the most effective prevention strategy for subjects at risk of atherosclerotic events.48


At the present time, measures of endothelial cell function and biocompatibility have advanced our understanding of the pathophysiology of atherosclerosis and its treatment. They are quickly becoming well-established surrogates of disease activity; however, the ideal test(s) of endothelial function have yet to be established. As we incorporate new biomarkers into global risk assessment, the endothelium is the logical target of study, given its unique position as both a sensor and participant in the atherosclerosis process. Recent evidence suggests that the mechanistic basis for the powerful predictive value of inflammatory biomarkers such as CRP may also reside at the level of the endothelium.16 Although endothelial function testing remains a research tool at the present time, it is our contention that this technology will figure prominently in risk assessment strategies in the future.


The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.