Adenosine does a lot of things. It's called a neuromodulator, because it modulates, or controls, the activity of neurotransmitter molecules including serotonin, norepinephrine, dopamine, and acetylcholine. The effect adenosine has depends on where in the body the adenosine is; adenosine can have opposite effects in different parts of the nervous system. Overall, adenosine acts as a tranquilizer, inhibiting nerve firing by inhibiting the release of excitatory neurochemicals. Centrally, in the brainstem and spinal cord, adenosine is a painkiller, but peripherally, in the outer reaches of the nervous system, it causes pain. Adenosine applied to the skin causes localized pain and vasodilation. Adenosine dilates blood vessels in the head and neck.

Caffeine, because it blocks adenosine, has the opposite effects of adenosine. Caffeine constricts blood vessels in the head and neck, and increases the release of excitatory neurochemicals, so increases the rate of nerve firing. That's why caffeine is stimulating.

Caffeine has only one known mechanism of action at likely doses: occupation and blockade of adenosine receptors. Whatever the cause of a given primary or secondary headache, if caffeine relieves that headache, it does so by blocking adenosine. That tells us something important about headaches relieved by caffeine: their causal mechanism involves adenosine binding to adenosine receptors.

Model of an adenosine molecule (green) bound to an adenosine receptor.

The concentration of adenosine in the head and neck is increased above normal during migraine episodes,1 so adenosine, which causes pain peripherally2 and is a vasodilator, readily explains the head pain and vascular dilation associated with migraine.

That caffeine oftentimes relieves migraine means adenosine must be a key part of the mechanism of migraine. This straightforward reasoning makes little impression on headache researchers, though they haven't completely ignored adenosine. In 1996 Professor Burnstock proposed that adenosine is the endogenous (originating within the body) cause of migraine pain.3 Always looking for new drug targets, headache researchers have investigated the possibility of using adenosine receptor agonists, molecules that bind to and activate adenosine receptors, as migraine medicines.*

The human nervous system compensates for caffeine's interference with adenosine by becoming more sensitive to adenosine. It does so by releasing more adenosine, increasing the number of adenosine receptors on the surfaces of neurons, increasing the affinity of those receptors for adenosine (increasing the amount of time adenosine molecules remain bound to adenosine receptors), and decreasing the rate at which adenosine is removed from the synapse (the gap between neurons) by transporter molecules that retrieve and ferry adenosine back inside neurons. All these changes tend to increase adenosine receptor activation, to compensate for adenosine receptors clogged by caffeine. Without that nervous adaptation the caffeine user would suffer nervousness, anxiety, restlessness, anorexia, sleeplessness, the jitters, muscle twitches, and nervous tics.

Absent caffeine, however, a caffeine user is oversensitive to adenosine; a caffeine user is adapted to exposure to caffeine, ill-adapted to its absence.

The caffeine paradox
The paradox of caffeine is that it both causes and relieves headache. The following excerpt, from a website offering information about headache, expresses the prevailing view of the relationship between caffeine and headache:

Q.: Is Caffeine a headache treatment or a headache trigger?

Caffeine can be both beneficial and harmful for a headache sufferer. Caffeine is a common ingredient in many prescription and over-the-counter headache medications. Caffeine additives make pain-relievers 40% more effective in treating headaches. Caffeine also helps the body absorb headache medications more quickly, bringing faster relief.

While caffeine-containing medications can be beneficial, these medications, combined with consuming too much caffeine (coffee, tea, soft drinks or chocolate) from other sources, may make you more vulnerable to getting rebound headaches.4

Interestingly, caffeine seems to both trigger and relieve migraines.5

Caffeine as risk factor for headache
This news bulletin was posted on the ACHEnet (American Committee for Headache Education) website:

A study reported in the December 2004 issue of Neurology looked at whether caffeine use increased the risk of chronic daily headache. The research group studied the caffeine usage of two groups of people in three metropolitan areas: a control group, whose members reported between two and 104 headache days per year, and another group whose members had more than 180 headache days per year. People in the upper 25 percent of those who consume caffeine in food or beverages, or who use an over-the-counter medication that contains caffeine, were defined as having "high caffeine exposure." The study concluded that both dietary and medicinal caffeine usage appear to represent only a modest risk factor for those getting chronic daily headache, regardless of the type of headache.6

The Neurology study described above, a type of study called a case-control study, fails to strike at the heart of the matter. The best way to determine caffeine's contribution to chronic headache would be to compare chronic headache in patients who use caffeine to chronic headache in patients who use no caffeine whatsoever. We would learn a great deal more, in other words, from a type of clinical study called a cohort study, in which one group of subjects is weaned from caffeine, another group kept on caffeine, and both groups followed along to see what happens to their headaches. That would be the equivalent of a clinical trial of caffeine abstinence as headache treatment. Such a trial would reveal, for the first time, what primary headache looks like absent caffeine.

The case-control study published in Neurology compares caffeine intake in subjects with episodic headache before they developed chronic headache to caffeine intake in subjects with episodic headaches who haven't developed chronic headache.7 A shortcoming of the study is that the estimates of caffeine intake rely on subjects' recollections of what they ate and drank, and which medicines they used, up to five years in the past. The biggest problem with the study design, however, is that only a very moderate caffeine intake is needed to cause headache. Withdrawal from a habit of only 100 mg caffeine per day can cause headache.8 The average person in the U.S. uses twice that much caffeine,9,10 so possibly all the subjects in the Neurology study use 100 mg or more caffeine per day. That means caffeine may have caused every headache in every subject in the study. Several subjects reported using no caffeine, but their caffeine abstinence wasn't verified by testing. Those subjects might be mistaken about their absolute lack of caffeine intake.

The gunshot wound analogy
Let's say a case-control study found that subjects with five or more gunshot wounds were slightly more likely than subjects with three or fewer gunshot wounds to develop gunshot death. The authors of that study might conclude that gunshot wounds are a modest risk factor for gunshot death.

All the Neurology study tells us is that among people with episodic headache, those who report heavy caffeine intake are more likely to develop chronic daily headache. If caffeine was the sole and indispensable cause of both episodic headache and chronic daily headache, the Neurology study wouldn't tell us so. A study of chronic daily headache patients who verifiably don't use caffeine might reveal that no such persons exist.

Caffeine use is so highly prevalent its effect on primary headache is nearly invisible, because primary headache is viewed against a nearly solid background of caffeine use. That may be why primary headache seems less disease than human attribute. If only half the population used caffeine, we would notice if everybody who didn't use caffeine was immune to primary headache. In reality, people who don't use caffeine are so rare their immunity to primary headache might escape our notice. We don't even know who those people are. Headache studies have consistently failed to establish a control group of caffeine nonusers to compare to users.

Caffeine as medicine
Many people find caffeine, taken in the form of coffee, tea, soda pop, or caffeine tablets relieves their headaches, and caffeine is included in many combination headache medicines. Unfortunately, caffeine isn't the headache panacea; it doesn't effectively abort all primary headaches in all patients. Why does caffeine so often abort primary headaches, and why does it so often fail to do so?

The vascular theory
There have been several theories that attempt to explain why caffeine relieves headaches. One is that caffeine works by constricting blood vessels. Blood vessels in the head and neck are dilated during a migraine episode, and caffeine does indeed constrict blood vessels in the head and neck. The conclusion that caffeine relieves headache by constricting blood vessels, however, is premised on the assumption that vascular dilation causes migraine headache. That assumption has it backwards: vascular dilation is a symptom of migraine, not the cause of migraine headache. Caffeine's headache-relieving power doesn't depend on its vasoconstrictive effect.

Oftentimes a migraine headache is a pounding headache, with pain spiking in time with the pulse. In some cases, dilated superficial temporal arteries visibly pulsate. Those symptoms led migraine theorists in the 1940s to believe vascular dilation and pulsation caused the associated headache pain. The vascular theory of migraine supposes vascular spasms cause migraine symptoms. According to the vascular theory, vascular constriction in the brain causes ischemia (inadequate blood flow causing tissue to be deprived of oxygen), which causes aura symptoms, followed by vascular dilation that causes the head pain. The brain itself is insensitive to pain, so the vascular theory postulates that migraine pain originates when dilated blood vessels in the meninges, the three layers of membrane surrounding the brain, press against or otherwise physically or chemically irritate sensory nerves.

The vascular theory has problems. One is that blood vessels routinely constrict and dilate without causing pain. Another is that the vascular constriction that happens during migraine episodes doesn't choke off cerebral blood flow enough to cause aura symptoms. Yet another is that experiments in which blood pressure in the brain was artificially increased failed to produce migraine-like symptoms. The central problem with the vascular theory, however, is that it assumes vascular changes cause migraine symptoms, when there are other logical possibilities: the headache could cause the vascular dilation, or a third, underlying factor could cause both the vascular dilation and the headache.

The vascular theory fails to take into account the neurochemical mechanisms that control vascular constriction and dilation. To understand the relationship between vascular dilation and headache, it's necessary to understand why blood vessels constrict and dilate. Blood vessels are surrounded by a layer of muscle cells that can contract to decrease the diameter of the vessel. The degree to which the muscle layer is contracted is referred to as vascular tone. The layer of muscle cells is under neurochemical control, so the diameter of blood vessels, or vascular tone, is under neurochemical control. The sympathetic nervous system controls blood vessel diameter, and norepinephrine is the final neurotransmitter of the sympathetic nervous system. Norepinephrine constricts blood vessels. Adenosine, by inhibiting the release of norepinephrine from neurons, causes vasodilation. Caffeine, by blocking adenosine, prevents adenosine from inhibiting norepinephrine release, so caffeine increases the amount of norepinephrine released by neurons, and thereby causes vasoconstriction.

The new scientific consensus is that migraine is a neurovascular disorder, the modified idea being that a central nervous system dysfunction leads to the vascular dilation. Many lay authors, however, still assume the vascular dilation directly generates migraine pain.

Adenosine is the likely culprit behind both the vasodilation and pain. Migraine is a derangement of normal neurochemical mechanisms, including those that control blood vessel diameter. During a migraine episode, the blood concentration of adenosine in the head and neck is elevated quite a bit above normal; by an average of 68 percent in one study.1 Adenosine causes pain and is a potent vasodilator, so it's not surprising that excessive adenosine in the head is associated with the head pain and vascular dilation that typically accompany migraine episodes. Nor is it surprising that caffeine, which blocks adenosine, remedies headache.

Another reason to believe adenosine is the endogenous (originating within the body) culprit in migraine: the drug dipyridamole, used to prevent stroke, works by inhibiting the reuptake and destruction of adenosine, thereby allowing adenosine to accumulate outside of cells. Not surprisingly, dipyridamole is notorious for causing headaches. In one study intravenously administered dipyridamole induced headache in all ten migraine patients, migraine in five of them, and induced headache in eight of the ten healthy controls, headache that qualified as migraine in one of those eight.144

Caffeine as adjuvant ingredient
Another theory is that caffeine in combination headache medicines works by enhancing the effects of the "hero" medicines it's combined with. One proposed mechanism for the adjuvant effect of caffeine is that caffeine expedites delivery of the other medicines by speeding up absorption of the hero medicines through the gut wall into the bloodstream, or by increasing heart rate, thereby speeding the hero medicines on their way.

The adjuvant theory doesn't explain why caffeine all by itself often effectively aborts headache, and has an analgesic effect on headache independent of other medicines.11 Moreover, because oftentimes caffeine all by itself aborts headache, the so-called adjuvant or synergistic effect of caffeine would be difficult to demonstrate. If a caffeine-containing combination headache medicine aborts a headache, it would be hard to determine how much of the credit goes to an adjuvant effect of caffeine rather than to the intrinsic ability of caffeine to abort headache. But caffeine does slow the rate at which aspirin and acetaminophen (paracetamol, Tylenol) are metabolized, so could make aspirin and acetaminophen more effective by keeping them in the body longer and at a higher concentration.

The real reason caffeine aborts primary headache
According to the prevailing view, caffeine aborts caffeine withdrawal headache episodes by reversing caffeine withdrawal, but aborts migraine by some different, unidentified mechanism.16 Yet caffeine has only one known mechanism of action at likely doses: occupation and blockade of adenosine receptors. Caffeine must, therefore, relieve both caffeine withdrawal headache and migraine by blocking adenosine receptors. Caffeine's ability to abort migraine and other primary headaches provides insight into the causal mechanism of those headaches: evidently, adenosine receptor activation is an essential part of the causal mechanisms of both migraine and caffeine withdrawal headache.

The caffeine paradox resolved
The mechanism of addiction and withdrawal readily explains the ability of caffeine to both cause and relieve headaches. Caffeine causes headaches by provoking neurochemical dependency and resulting susceptibility to caffeine withdrawal symptoms including headache, and relieves those symptoms by reversing caffeine withdrawal.

Adenosine and secondary headaches
During the 2000 Republican National Convention, former President Gerald Ford went to the hospital, complaining of sinus pain. He was treated for a headache and released. The next morning he was back at the hospital. This time his condition was correctly diagnosed: he had suffered a mild stroke, possibly two.

Why did stroke give Ford a headache? Because strokes cause the release of adenosine, and adenosine causes headache. Strokes cause ischemia, thereby hypoxia (shortage of oxygen), in the brain. During hypoxia, neurons release adenosine in an attempt to restore oxygen homeostasis, the balance between demand for and supply of oxygen. Adenosine inhibits the release of excitatory neurotransmitters, so damps nerve firing and thereby decreases oxygen demand. At the same time, adenosine also dilates arteries in the head and neck, in an attempt to keep enough blood and oxygen flowing to the brain to maintain consciousness. But adenosine has other effects besides vasodilation: it also causes pain by binding to adenosine receptors on sensory neurons in the head. During Ford's stroke the adenosine released by his nervous system in an attempt to restore oxygen homeostasis gave him a headache.

All headaches, primary and secondary, whatever their cause, share a portion of their causal mechanism: all involve adenosine and adenosine receptor activation. Hypoxia promotes migraine.17 Hypoxia also causes the headache that typically follows carbon monoxide poisoning (assuming the victim survives). Headache following head trauma may be due to the release of adenosine triphosphate (ATP)-rich cytoplasm from ruptured cells. ATP, once outside of cells, is rapidly converted to adenosine. ATP and adenosine released from damaged tissues may also cause the sedation and other symptoms of shock following trauma.

Caffeine alters the nervous system
Caffeine use causes a wide range of long- and short-lasting adaptive changes to the nervous system, changes that render the nervous system dependent on caffeine and set the nervous system up for a disturbance when it goes too long without it.

Caffeine increases the affinity and density (number) of type A118 and type A2A adenosine receptors.19 A migrainer is excessively sensitive to adenosine and excessively insensitive to excitatory neurotransmitters. These are exactly the adaptations caffeine causes. Caffeine, in other words, alters the nervous system in just the right way to make a person a migrainer. No wonder, therefore, that exposure to caffeine can be followed by a withdrawal syndrome indistinguishable from a migraine episode.

In experiments, rats given water spiked with caffeine for two weeks had nearly ten times more adenosine in their blood plasma than had controls given plain water. Intravenous caffeine also increased plasma adenosine concentration in rats.148

When the caffeine-spiked water was replaced in the evening with plain water, the next morning plasma concentration of adenosine in the formerly caffeine-drinking rats had decreased to less than half that of controls never given caffeine. This doesn't seem to jibe with the increased blood concentration of adenosine in humans during migraine episodes. But rats metabolize caffeine much more quickly than humans do: the half-life of caffeine in rats averages about one hour, compared to three to seven hours in humans.150

When the usual regular dose of caffeine is delayed, the caffeine remaining in the body is catabolized (dismantled and thereby deactivated) by enzymes in the liver. With caffeine gone, formerly blocked adenosine can access, bind to, occupy, and activate now unblocked and excessively numerous adenosine receptors. With too many adenosine molecules attached too long to too many adenosine receptors, the nervous system malfunctions.

Excessive adenosine, combined with excessive sensitivity to adenosine, generates the pain of caffeine withdrawal headache and likely causes migraine head pain as well. In a person abnormally insensitive to excitatory neurotransmitter chemicals inhibited by adenosine, excessive adenosine receptor activation causes a shortfall in excitatory neurotransmitter receptor activation. As a result, systems controlled by those neurotransmitters malfunction, and those malfunctions generate migraine symptoms. A shortage of serotonin receptor activation prevents serotonin from appropriately blocking pain. A shortage of acetylcholine, a neurotransmitter essential to vision, in the retinas may cause the visual disturbances associated with migraine. A shortage of dopamine may cause the emotional and behavioral effects associated with migraine. And a shortage of norepinephrine, the final neurotransmitter of the sympathetic nervous system, causes the symptoms of sympathetic hypofunction (inadequate functioning of the sympathetic nervous system) associated with primary headache.

High prevalence of migraine
Most people regularly use caffeine, but the exact percentage of the general population that uses caffeine isn't known. Among a sample of 14 studies of caffeine use in various subpopulations including premenopausal women,20 Vermonters,21 retired military,22 twins,23 the general population,24 female nurses,25 workers,26 women in the first trimester of pregnancy,27 people over 50,28 couples of childbearing age,29 and children,30 caffeine use averages 85 percent; the median is 86.9 percent. That figure matches the result of a 2005 survey study of the U.S. general population in which 87 percent of 18,081 subjects reported dietary caffeine intake31 and 71 percent said they drank coffee.

The actual prevalence of caffeine use is very likely higher, as all but one of those studies are survey studies, and subjects in survey studies very likely underreport caffeine use, due to what is called recollection bias. Moreover, all of those survey studies leave out commonplace sources of caffeine. The survey studies ask subjects about dietary sources of caffeine only, so don't include medicinal sources. In some of the studies subjects were asked only about their coffee use. In some of the studies subjects were asked about colas but not about other flavors of caffeinated soda, and in a few of the studies weren't asked about caffeinated sodas at all.

The study of caffeine use prevalence among children was a diary study, rather than a survey study, so may more accurately represent caffeine use prevalence in the general population. In that study, 98 percent of the child subjects used caffeine during the week-long study period.30

We know for sure that somewhere between 87 and 100 percent of the general population of the U.S. consumes caffeine at least occasionally. That means caffeine withdrawal could account for the near-universal lifetime prevalence of headache among some groups. A study of Danish adults found 88 percent of women and 69 percent of men have tension-type headaches some time in their lives, and 93 percent of men and 99 percent of women have some type of headache some time in their lives.32 One study found a 98.1 percent lifetime prevalence of headache among 1208 college students in Singapore.33 Among a group of 96 medical students in Spain, lifetime headache prevalence was 95.8 percent; 96.6 percent among the women.34 Among 403 medical students in Oman, lifetime prevalence of headache was 98.3 percent.35

The high prevalence of headache among college students, and especially among medical students, may be due to their heavy use of caffeine as a study aid. Migraine prevalence roughly correlates to national per capita caffeine intake. Migraine is highly prevalent in Germany,36 Switzerland,37 Denmark,38 the United Kingdom,39 Canada,40 France,41 and the United States,42 countries with high estimated per capita caffeine consumption.

Migraine and natural selection
Some authors have wondered why migraine is so common, have wondered why natural selection hasn't ruthlessly weeded it out, and have speculated that migraine might confer some reproductive advantage that outweighs its disadvantages.43 It's hard to imagine, though, how being episodically disabled could help a parent produce healthy offspring. Nevertheless, it may be caffeine use confers nutritional or behavioral advantages that more than compensate for the headaches it causes. Caffeine, a potent stimulant, may increase productivity and therefore wealth and concomitant good nutrition and health care. Coffee, a major source of caffeine, is also a major source of niacin, a necessary nutrient, and coffee also contains antioxidants. Coffee drinking is associated with a lower risk of type 2 diabetes,44 and protects against cardiovascular disease.45 Green tea protects against various cancers. Boiling water to brew coffee or tea sterilizes the water and thereby protects against waterborne infectious diseases.

Though nearly everybody these days is regularly exposed to caffeine, on an evolutionary time scale humankind was introduced to caffeine only recently. Primary headache including migraine may be an artifact of our incomplete adaptation to caffeine.