There are probably several reasons caffeine, despite possessing all the qualities needed to cause the entire phenomenon of primary headache, is considered exculpated, when it hasn't been. Chief among them are the peculiarities of caffeine, which make it difficult to connect caffeine to the headaches it causes.

The long delay between exposure to caffeine and onset of headache obscures the causal relationship.

Headache doesn't necessarily follow caffeine intake, an inconsistency that further obscures the connection. If a dose of caffeine was reliably followed by a headache, the association between the two would stand out more clearly. Exposure to caffeine, however, isn't sufficient to cause headache; a caffeine withdrawal headache also requires a delay in caffeine intake.

Another peculiarity of caffeine is that it often but not always relieves primary headache episodes. The failure of caffeine to abort headaches in some instances or cases may have convinced many patients and physicians that those particular instances or cases of headache cannot be due to caffeine withdrawal. Headache researchers go even further and conclude that caffeine withdrawal can't be the cause of any primary headache.

The near-universal prevalence of both caffeine use and primary headache effaces the relationship between the two, both seeming to blend into the background of the human condition.

Entrenched migraine mythologies, including the heterogeneity, vascular, stress, and trigger theories, relegate caffeine to a minor role in primary headache. The heterogeneity theory muddies the waters and implies each type of primary headache has a different causal mechanism. The vascular theory mistakes a symptom for a cause. The stress theory misinterprets caffeine-caused anxiety as an abstract force generated by everyday life. The trigger theory says anything and everything can precipitate a headache, so caffeine gets lost among a crowd of putative triggers including oranges, changes in barometric pressure, hats, anger, aspartame, bacon, computer monitors, Doritos, whipped cream, poor posture, and perfume.

By the 1970s, when it was discovered that caffeine works by blocking adenosine, the assumption that caffeine doesn't cause migraine had ossified into doctrine. The vascular theory prevailed then, including the idea that caffeine relieved headache by constricting blood vessels, and that caffeine withdrawal headaches were due to "rebound" vasodilation when the caffeine wore off. Nobody, it seems, connected caffeine's newly-discovered mechanism of adenosine antagonism with caffeine's vasoconstrictive effect, with caffeine addiction, or with caffeine's ability to relieve both primary headache and caffeine withdrawal headache.

Caffeine is addictive, and addiction very often trumps reason or bends it to its purposes; if reason could easily overcome addiction, addiction wouldn't be a problem. People don't seem to be able to think straight when it comes to caffeine. Obviously, primary headache patients shouldn't use an elective, headache-causing drug, yet most—perhaps all—do.

Because they want to keep using caffeine, many patients choose to believe caffeine cannot be the cause of their migraines. They blame their headaches on things never demonstrated to cause headache, or on unidentified causes, and reject the possibility that caffeine, a drug demonstrated to cause headache and which they regularly use, might be the culprit. Many other primary headache patients are either unaware they use caffeine or have convinced themselves they don't use it.

Although many people who get caffeine withdrawal headaches understand what causes them, it's highly likely many others don't realize their headaches are caffeine withdrawal headaches. Their headaches may be a mystery to them, or they may mistakenly blame them on foods, stress, lack of sleep, the weather, odors, or other things. It must be physicians diagnose at least some cases of caffeine withdrawal headache as one or another primary headache.

Many people report having caffeine withdrawal headaches as well as migraine episodes, and claim they can clearly tell the difference. In such cases patients may have mistaken one variable condition for multiple conditions. Patients may correctly attribute their mild or moderate headaches to caffeine withdrawal, but fail to recognize their severe caffeine withdrawal headaches as such, and instead consider them mysterious, or attribute them to other causes.

Misunderstandings and misinterpretations of that kind are possible because patients can inadvertently and unknowingly ingest caffeine, and, just as important, can inadvertently and unknowingly delay regular caffeine intake. Physicians err whenever they assume caffeine can't be the cause of migraine in a patient who reports using caffeine. They err as well when they assume caffeine can't be the cause of migraine in a patient who reports using no caffeine, as patients can consume caffeine without being aware of it.

Headache researchers are also physicians with their own patients. The personal relationship between physician and patient, vital to a successful medical practice, works against physicians in their role as headache researchers, because it biases their thinking. Their patients trust and respect them, and the physicians return that trust and respect. This reciprocal trust and respect between doctor and patient becomes a disadvantage in headache research. When primary headache patients report consuming no caffeine whatsoever, headache physician/researchers tend to assume such testimony is reliable, and accept it at face value.

Headache theorists reason that, because many patients report caffeine abstinence, caffeine must not cause any primary headaches. Yet most primary headache patients report using caffeine, so in most cases caffeine could be the culprit. Moreover, primary headache patients who report using no caffeine may be mistaken.

Pharmaceutical corporations sell billions of dollars worth of headache and migraine medicines each year, including tons of caffeine sold in the form of headache remedies. The companies that manufacture and market caffeine-containing headache remedies don't want any discussion of the possibility their wares may cause more headaches than they relieve. Those companies blame their customers for chronic headaches caused by addiction to their caffeine-containing headache medicines, claiming it's their customers' fault for failing to follow the instructions on the label and taking the medicines too often. But if a company sells an addictive drug that many of its customers become addicted to, whose fault is that? Pharmaceutical companies would prefer migraine remain a chronic, incurable, but manageable condition—managed by steady use of their increasingly sophisticated and expensive proprietary migraine drugs. If research determined caffeine withdrawal was the major or sole cause of primary headaches including migraine, and that most or all migraine patients can, therefore, eliminate their migraines without resorting to expensive drug therapies, companies that make and market migraine medicines stand to lose a lot of money.

Food companies that market caffeine don't want consumers making connections between dietary caffeine and headaches. That's why four major makers of caffeine-containing beverages funded a trial study of caffeine withdrawal. The study was funded by the North American Branch of the International Food Information Council, an industry front. The IFIC Technical Committee on Caffeine consists of the Coca-Cola Company, Kraft Foods, Inc. (Maxwell House Coffee, General Foods International Coffees), the Proctor & Gamble Company (Folger's Coffee), and PepsiCo. Several articles describing the results were published in medical journals. The authors of the articles conclude that caffeine withdrawal is uncommon, therefore insignificant, and that reported caffeine withdrawal symptoms are largely due to expectation, that is, to the placebo effect, rather than to the pharmacological effects of caffeine.130,131

Their conclusions contradict the body of earlier research on caffeine withdrawal. In previous studies, most subjects deprived of caffeine got a headache. In one study, all seven subjects abruptly denied a daily dose of 100 mg caffeine—about half the estimated U.S. per capita daily caffeine intake9,10—suffered headache. Five of the seven suffered severe headache, and three vomited.8 In another trial study, moderate-to-severe headache among 64 normal subjects denied their usual caffeine intake for 48 hours increased from three to 52 percent.132 The IFIC-sponsored authors, however, claim their study trumps the contrary findings of previous studies because those earlier studies were smaller and used flawed methodology.

In the Dews trial, subjects were divided into three groups, given packets of instant coffee, and instructed to use each packet on a specified day. The trial was blinded: subjects were not told the true object of the study, but led to believe the study was mainly about the flavor of the coffee, and were not told that some of the packets contained decaf coffee. One group received regular coffee throughout the trial period. Another group received decreasing amounts of caffeine each day, so that caffeine was gradually withdrawn. The third group got regular coffee for the first three days then decaf coffee for the next five days. Six of eighteen subjects in the abrupt withdrawal group reported withdrawal symptoms such as headache and tiredness. That's 33.3 percent, lower than the 52 percent and 100 percent in the previous studies described above, and lower than the 50 percent average in all relevant studies including the Dews study. The Dews authors conclude the higher numbers in the previous studies resulted from expectation, also known as the placebo effect.

The irony is that the Dews study protocol induces the placebo effect. Subjects in the Dews study expected regular coffee but sometimes received decaf, so the decaf functioned as placebo caffeine. Genuine caffeine and placebo caffeine have similar effects. The authors of a 2004 study of this phenomenon conclude:

The results indicate that caffeine expectation induces dopaminergic placebo effects, and that these effects are similar to previous findings with oral caffeine. The results therefore suggest that caffeine and placebo caffeine may share some dopaminergic mechanisms of action.133

Because expectation has a profound effect on headache, the expectation designed into a trial study can minimize or maximize the number and severity of headaches subjects get during trial studies. The Dews authors designed a trial that may have suppressed headache with caffeine placebo. Previous study designs may have promoted headaches by inducing expectation of caffeine withdrawal. Nevertheless, the conclusion of the Dews study authors, that caffeine withdrawal headache is uncommon and due largely to expectation rather than to the pharmacological effects of caffeine, amounts to wishful thinking on the part of their corporate sponsors. It's corporate propaganda masquerading as science.

What we want to know is how prevalent, frequent, and severe caffeine withdrawal headache is among the general population. It may be no trial study can duplicate the real-world circumstances surrounding caffeine withdrawal. In a non-blinded trial study, subjects deprived of caffeine are aware they are deprived, and in a blinded trial study, subjects deprived of caffeine must be given placebo caffeine. In the real world, in contrast, a caffeine user can be deprived of caffeine unawares without receiving placebo caffeine. In that situation the caffeine user neither expects a caffeine withdrawal headache nor is protected from caffeine withdrawal headache by caffeine placebo. So the expectations that influence headache may be unavoidably different for subjects participating in scientific trials than for the typical caffeine user.

In any case, there's another, more important factor the Dews researchers didn't take into account: subjects may have been getting caffeine from other sources besides the coffee packets supplied by the researchers. Even a small amount of extracurricular caffeine could have prevented withdrawal symptoms. The researchers instructed the subjects to avoid outside sources of caffeine, and assumed the subjects complied, but the subjects may have unknowingly failed to comply. It's not known how likely a subject instructed to avoid caffeine will manage to successfully avoid caffeine.

Primary headache is highly prevalent and a significant health problem, so until primary headache is demonstrated to be something apart from caffeine withdrawal headache, the Dews authors can't soundly conclude that caffeine withdrawal symptoms are uncommon and insignificant. In any event, the Dews study results don't overturn all previous research. Dr. Griffiths, long-time investigator of caffeine withdrawal, rebuts the Dews authors with a comprehensive review of the relevant research to date:

A literature search identified 57 experimental and nine survey studies on caffeine withdrawal that met inclusion criteria. The methodological features of each study were examined to assess the validity of the effects. RESULTS: Of 49 symptom categories identified, the following ten fulfilled validity criteria: headache, fatigue, decreased energy/activeness, decreased alertness, drowsiness, decreased contentedness, depressed mood, difficulty concentrating, irritability, and foggy/not clearheaded. In addition, flu-like symptoms, nausea/vomiting, and muscle pain/stiffness were judged likely to represent valid symptom categories. In experimental studies, the incidence of headache was 50 percent and the incidence of clinically significant distress or functional impairment was 13 percent. Typically, onset of symptoms occurred 12 to 24 hours after abstinence, with peak intensity at 20 to 51 hours, and for a duration of two to nine days. In general, the incidence or severity of symptoms increased with increases in daily dose; abstinence from doses as low as 100 mg/day produced symptoms. Research is reviewed indicating that expectancies are not a prime determinant of caffeine withdrawal and that avoidance of withdrawal symptoms plays a central role in habitual caffeine consumption.134

The same sort of thing happened with gastric ulcer. As with migraine, doctors long blamed gastric ulcer on diet and stress. Medicines used to manage ulcer were best sellers, and at one point in the 1960s gastric resection accounted for one in four major surgeries. There was grumbling when Barry Marshall discovered the cause of and cure for gastric ulcer. Marshall destroyed the ulcer industry when he demonstrated that gastric ulcer can be cured—not managed but cured—with antibiotics. A lot of people who had been making lots of money managing gastric ulcer had to find another way to earn a living.

The logical strategy for sorting out the cause of medicine overuse headache would be to determine the contribution of each individual medicine by observing medicine overuse headache absent each medicine. A recent Danish study, Zeeberg, et al.,135 did just that, separating out the effect on medicine overuse headache of quitting particular headache medicines for two months. Subjects were divided into eight groups according to the type of headache medicine they had overused. The eight categories of headache medicine the subjects had overused are listed in this table from the article describing the study:

Inexplicably, the word caffeine doesn't appear in this table, nor anywhere else in the article. That's quite odd, given that medicines associated with medicine overuse headache are often administered in combination preparations that include caffeine, and that caffeine is a demonstrated cause of headache8,132 and a risk factor for chronic headache,7 so may play a causal role in medicine overuse headache. By pretending medicinal caffeine doesn't exist, the authors avoid illuminating that role. I won't say something is rotten in the state of Denmark, but something weird is going on; it's as if an invisible force keeps headache investigators from focusing on caffeine. A similar study that included caffeine abstinence would be worthwhile.

Medicine overuse headache was formerly called medicine rebound headache. One feature of medicine overuse headache is that abrubt discontinuation of the overused medicine induces headache episodes. So-called "rebound" headache, therefore, is clearly withdrawal headache; the substitution of the word "rebound" for the more descriptive and accurate word, withdrawal, seems a perhaps unconscious attempt to avoid invoking the mechanism of addiction and withdrawal. One author refers to medicine overuse headache as "rebound-withdrawal headache."145

As soon as medicine overuse headache is acknowledged to be withdrawal headache, a logical process of elimination quickly leads to caffeine: Of the medicines associated with medicine overuse headache, only butalbital, ergotamine, caffeine, and codeine are known to be addictive. Of those four, only caffeine is a demonstrated cause of headache.

Moreover, there's no demonstrated example of any headache medicine being associated with chronic headache, medicine overuse headache, "rebound" headache, or "rebound-withdrawal" headache in a patient who doesn't use caffeine. It may be all people who verifiably use no caffeine are immune to those types of headache.

There shouldn't be any mystery about what causes primary chronic headaches, given the results of a 2003 study in which quitting cola eliminated chronic headache in 100 percent of subjects. 36 children and adolescents with daily or near-daily headache who consumed from 193 to 386 mg caffeine per day in the form of cola were encouraged to wean themselves from cola. Headaches totally ceased in 33 (91 percent) of the subjects. Daily or near daily headaches ceased in the remaining three subjects, though they continued to get infrequent migraines without aura.140

Excedrin headache
On Excedrin TV commercials caffeine is coyly referred to as "an enhancing ingredient." On the Excedrin website caffeine is called "therapeutically active caffeine™."

The maker of Excedrin paid prominent headache researchers to help secure FDA approval of a new product called Excedrin Migraine.136 Excedrin Migraine is not a new medicine developed specifically for migraine. Rather, it is a pre-existing product, Extra-Strength Excedrin, with a different label; both products contain 250 mg acetaminophen, 250 mg aspirin, and 65 mg caffeine.

Excedrin Migraine was created for marketing purposes. One purpose of Excedrin Migraine is to increase shelf space for Excedrin brand products. This is known as shelf space marketing, the struggle for limited space on store shelves, especially the coveted shelves at or near eye level. Giving Extra-Strength Excedrin two different names doubles the amount of shelf space the product occupies, thereby shoving competing brands onto less desirable shelves, or, ideally, off a store's shelves altogether.

More recently, the maker of Excedrin added Excedrin Tension Headache to its product line. Excedrin Tension Headache is similar to Extra-Strength Excedrin and Excedrin Migraine, except it has no aspirin and double the acetaminophen. Apparently the makers of Excedrin figured they could get away with different labels on two identical medicines but not three. Or perhaps they realized that offering the exact same formula to treat both tension-type headache and migraine would have undermined the premise that the two headache types are separate and distinct in the first place. So they altered the formula of the third version. Most likely, Extra-Strength Excedrin, Excedrin Migraine, and Excedrin Tension Headache are all equally effective at relieving either migraine or tension-type headaches.

There is also Excedrin Sinus Headache, containing acetaminophen and phenylephrine* but no caffeine. Headache researchers now understand that most so-called sinus headaches are migraine headaches. In one recent study, 88 of 100 subjects with self-diagnosed sinus headaches actually have primary headaches. 86 had migraine or probable migraine. Only 3 had headaches secondary to rhinosinusitis.141 So the makers of Excedrin are invested in the multiple disease theory of primary headache even when it contradicts the prevailing view, and exploit the theory to literally widen their product line and muscle their competition off store shelves.

One influential researcher who helped get Excedrin Migraine approved told me he believes migraine and tension-type headache are distinct and separate entities. He believes caffeine-containing headache medicines such as Excedrin aren't particularly culpable in medicine overuse headache. He believes addiction to Excedrin is an insignificant problem. He concludes, based on the results of a case-control study, that caffeine is only a modest risk factor for chronic headache. He's paid for thinking along those lines. Members of congress may insist campaign contributions don't influence them, but scientists know better, or should know better. He refuses to do a cohort study, which, admittedly, would be much more difficult and expensive than the Neurology case-control study, but would, unlike the case-control study, reveal the role caffeine plays in primary headache. Who would fund such a study, anyway? Not the makers of Excedrin.

Headache is the complaint neurologists hear most often from their patients. Rarely, headache indicates life-threatening illness, but the overwhelming bulk—90 percent—of headaches reported to neurologists and other physicians are primary headaches such as migraine, which, though painful, distressing, and unpleasant, seem to do little permanent physical damage. For a long time, primary headaches were a waste of time for physicians, who couldn't do much about them. Besides, primary headaches are self-limiting; a surefire treatment for migraine is to simply wait until the episode resolves itself, and, voila, the patient is cured—at least until the next episode.

That state of affairs went on for a long time, until the introduction of drugs, including aspirin and ergot alkaloids, effective at aborting migraine episodes. It wasn't until the triptan medicines were developed in the early 1990s, however, that migraine became really big business. And what a business: the market potential is enormous. An estimated 12 to 15 percent of the U.S. population currently has migraine. That's 36 to 45 million potential customers in the U.S. alone. Assuming an average of three migraine episodes per month per person, multiplied by 45 million migrainers, that's over a billion and a half migraine episodes a year. If every one of those migraine episodes was treated with a migraine abortant pill costing $10, that would add up to U.S. sales of over $15 billion per year, year after year. And that's just migraine. It's thought that around 80 percent of the general population of the U.S. suffers tension-type headaches.

Headache-specific medical journals are larded with estimates of how much migraine goes undiagnosed and untreated. Such articles amount to sales tools for pharmaceutical companies, used to illustrate to potential investors the untapped market potential of migraine drugs. Headache journals also publish estimates of worker productivity lost to migraine, articles intended to justify the cost of migraine treatments to the empoyers and governments that pay for most of them. So there's plenty of research on behalf of the bottom line.

Headache researchers show no interest, however, in investigating why and how caffeine relieves migraine. Caffeine and caffeine withdrawal headache appear to readily explain nearly every symptom and aspect of primary headache, yet headache researchers refuse to look into the relationship between caffeine withdrawal headache and primary headache. Nobody has investigated the effect of caffeine abstinence on primary headache, mainly because there's no money in it. Caffeine abstinence isn't a product that can be sold at a profit.

Researchers consistently neglect to rule out caffeine withdrawal headache during studies of migraine and migraine treatments. Caffeine withdrawal is possible in any person using caffeine, so the only known way to rule out caffeine withdrawal is to rule out caffeine intake, yet researchers allow subjects participating in migraine studies to freely consume unmeasured amounts of caffeine. Presumably, any severe caffeine withdrawal headache that occurs during migraine studies is counted as a migraine. As a result, migraine and caffeine withdrawal headache remain confounded. The working assumption in migraine research is that the two conditions are distinct and separate, yet, astoundingly, researchers have never bothered to separate them.

We know caffeine can abort both primary headache and caffeine withdrawal headache. The effectiveness of triptan medicines at aborting primary headaches has been the focus of many trial studies. Researchers haven't, however, investigated the effect of triptans on caffeine withdrawal headache. Perhaps triptans abort caffeine withdrawal headaches as effectively as they abort migraine headaches. Perhaps some or all headache episodes aborted by triptans are caffeine withdrawal headaches. We simply don't know, because caffeine and caffeine withdrawal headache haven't been controlled for during studies of triptan medicines.

All the parties involved—headache patients, their physicians, food and drug companies, headache researchers, and medical journals—are disinclined to investigate the relationship between caffeine and migraine. Headache patients addicted to caffeine don't want to believe caffeine is the problem. Physicians don't want to contradict and alienate caffeine-addicted headache patients who insist caffeine cannot possibly cause their migraines. Pharmaceutical and food corporations are financially motivated to downplay the role of caffeine in migraine. The reasoning of headache researchers and headache journal editors is bent toward the interests of the pharmaceutical companies funding them. This interlocking web of motivations, a powerful confluence of drug addiction and money, may not constitute a formal conspiracy, but nonetheless effectively suppresses investigation into the role caffeine plays in migraine.

Defective diagnostic criteria
The prevailing but undemonstrated assumption that primary headache can occur absent caffeine withdrawal has been integrated into the International Headache Society (IHS) diagnostic criteria for migraine. Because of this, many headache researchers and physicians may believe caffeine withdrawal has been ruled out as a possible cause of migraine. It hasn't been. Rather, the IHS criteria for migraine are defective.

ICHD-ll diagnostic criteria for migraine without aura At least five attacks fulfilling criteria B, C, and D   Headache attacks lasting four to 72 hours (untreated or unsuccessfully treated)   Headache has at least two of the following characteristics: unilateral location pulsating quality moderate or severe pain intensity aggravation by or causing avoidance of routine physical activity (e.g., walking or climbing stairs) During headache at least one of the following: Nausea and/or vomiting photophobia and phonophobia Not attributed to another disorder

Caffeine withdrawal headache often fulfills criteria A, B, C, and D, but cannot possibly fulfill criterion E, which requires the physician rule out conditions including caffeine withdrawal headache before diagnosing migraine. By definition, therefore, migraine cannot be caffeine withdrawal headache, nor vice-versa. Yet no physician or researcher has ever ruled out caffeine withdrawal headache before diagnosing migraine. Criterion E separates by decree two conditions never separated in fact.

The IHS diagnostic criteria for caffeine withdrawal headache are defective as well:

ICHD-ll diagnostic criteria for caffeine-withdrawal headache Bilateral and/or pulsating headache fulfilling criteria C and D   Caffeine consumption of at least 200 mg/day for more than two weeks, which is interrupted or delayed   Headache develops within 24 hours after last caffeine intake and is relieved within one hour by 100 mg caffeine   Headache resolves within seven days after total caffeine withdrawal

It hasn't been demonstrated that caffeine withdrawal headache is never unilateral, so criterion A is wrong. Criterion B formerly specified caffeine consumption of at least 500 mg/day, so this latest revision is an improvement. Nevertheless, the average person fulfills criterion B: one study calculated that the average person in the U.S. consumes 211 mg/day caffeine;10 another study concludes the average caffeine user in the U.S. ingests 193 mg/day.11 Withdrawal from 100 mg/day caffeine has been demonstrated to cause severe headache and nausea/vomiting.8 There's no demonstrated evidence two weeks of daily caffeine intake are required before a caffeine withdrawal headache can happen; withdrawal from a single, isolated dose of caffeine may be sufficient. Both requirements of criterion C are wrong: caffeine withdrawal headache can occur more than 24 hours after the latest dose of caffeine, and it hasn't been demonstrated that administration of 100 mg caffeine always relieves caffeine withdrawal headache. Criterion D is correct but neglects to specify how total withdrawal from caffeine is to be confirmed.

When caffeine withdrawal headache fulfills diagnostic criteria A through D for migraine, migraine can't be ruled out. Conversely, when migraine fulfills corrected diagnostic criteria for caffeine withdrawal headache, caffeine withdrawal headache can't be ruled out in caffeine users. Physicians who claim they can distinguish between the two conditions in a caffeine user can't explain how they are able to do so.

Nobody knows what percentage of migraine caffeine causes. It could be zero percent, 100 percent, or any percent in-between. The correct figure is very likely greater than zero percent, because caffeine withdrawal headache often closely resembles migraine, so must sometimes be mistaken for migraine. How often this happens, nobody knows. Neither does anybody know whether some or all caffeine withdrawal headaches are genuine migraine, or merely resemble migraine. In any case, caffeine withdrawal headaches, including those rightly or wrongly diagnosed as migraine or other primary headache, can be eliminated by successfully avoiding caffeine. For some or perhaps all migraine patients, therefore, caffeine abstinence would be a safe, cost-free, and effective treatment. Unfortunately, the effectiveness of caffeine abstinence as a treatment for migraine hasn't been investigated.

What patients should do
All primary headache patients should try caffeine abstinence. After all, it's only common sense that a headache patient should avoid elective, headache-causing drugs. Granted, nobody knows how effective caffeine abstinence is against headache, as the requisite study hasn't been done. Caffeine abstinence as a headache therapy may be like sexual abstinence as a birth control method: theoretically, sexual abstinence should be 100 percent effective at preventing pregnancy, but in practice is only 70 percent effective. The problem, of course, is patient noncompliance.

Patients who try caffeine abstinence and find it doesn't help shouldn't conclude caffeine withdrawal doesn't cause their headaches. Neither patient nor physician can soundly conclude so without testing the patient's saliva or other body fluid to verify the patient is truly getting no caffeine, as caffeine could be getting past them without them being aware of it. If a patient tries caffeine abstinence and it doesn't work, the patient and physician shouldn't give up on this approach. The logical strategy is to assume caffeine causes the patient's primary headaches, act as if the patient is getting caffeine from some source, and keep trying to identify that source.

Caveat: patients shouldn't quit caffeine abruptly, as that may induce withdrawal headaches. Instead, patients should gradually wean themselves from caffeine over several days.

Patients can expect that attempts to wean themselves from caffeine will alter the frequency, severity, pattern, or character of their headaches. Headaches may initially worsen, but in the long run will likely improve. Even if headaches don't cease, the changes indicate caffeine plays an important role, and should encourage patient and physician to continue focusing on caffeine as the most likely culprit.

As they reduce their caffeine intake, patients will probably find that headache medicines, including caffeine, become more effective. Patients may find, for example, that ibuprofen or naproxen sodium will effectively relieve their headaches, even if those medicines hadn't previously worked for them. So there are several good reasons to attempt caffeine abstinence.

What physicians should do
Physicians should instruct all primary headache patients to wean themselves from and stop using caffeine. If headaches continue, the patient should be tested to verify caffeine abstinence.

What researchers should do
First things first: migraine should be separated from caffeine withdrawal headache. Until migraine absent caffeine use has been demonstrated, theorizing about other causes for migraine is premature.

Headache researchers should perform a cohort study in which migraine patients are instructed to wean themselves from and abstain from caffeine. Subject compliance must be verified by sufficiently frequent testing.

In all studies of migraine and migraine treatments henceforth, caffeine withdrawal headache should be ruled out by instructing subjects to wean themselves from and abstain from caffeine, and testing their saliva or other body fluid to verify compliance