The clinician's first goal when confronted with a patient with cranial pain is to determine if the headache is primary or secondary. The main primary headache syndromes are migraine, tension-type headache, cluster headache, or one of the trigeminal–sympathetic migraine variants of migraine or cluster. These tend to be chronic, recurrent, and unattended by other symptoms and signs of neurologic disease. Familiarity with the variety of symptoms, temporal profiles, and accompanying features of the primary headache disorders, and the proclivity
for most of them to be familial, assist in identifying them from the patient's description. There should be little difficulty in recognizing the secondary headaches of diseases such as glaucoma, purulent sinusitis, subarachnoid hemorrhage, and bacterial or viral meningitis provided that these sources of headache are kept in mind. A fuller account of these types of "secondary" headache syndromes is given in later chapters of the book, where the underlying diseases are described. All other headaches that by their localization, quality of pain, and precipitating characteristics do not conform to one of the primary types should be suspected of being symptomatic of a cranial, cervical, or systemic disorder. Nonetheless, in many instances no such underlying cause will be found after investigation.
Migraine is a highly prevalent and largely familial disorder characterized by periodic, commonly unilateral, often pulsatile headaches that begin in childhood, adolescence, or early adult life and recur with diminishing frequency during advancing years.
Two closely related clinical syndromes have been identified, the first called migraine with aura and the second, migraine without aura (terminology of the International Headache Society). For many years, the first syndrome was referred to as classic or neurologic migraine and the second as common migraine. The ratio of classic to common migraine is 1:5. Either type may be preceded by vague premonitory changes in mood and appetite. Migraine with aura is ushered in by a disturbance of nervous function, most often visual, followed in a few minutes to hours by hemicranial (or, in about one-third of cases, bilateral) headache, nausea, and sometimes vomiting, all of which last for hours or as long as a day or more. Migraine without aura is characterized by an unheralded onset over minutes or longer of increasing hemicranial headache or, less often, by generalized headache with or without nausea and vomiting, which then follows the same temporal pattern as the migraine with aura. Sensitivity to light, noise, and often smells (photophobia, phono- or sonophobia, and osmophobia) attends both types, and intensification with movement of the head is common. If the pain is severe, the patient prefers to lie down in a quiet, darkened room and tries to sleep. The hemicranial and the throbbing (pulsating) aspects of migraine are its most characteristic features in comparison to other headache types. Each patient displays a proclivity for the pain to affect one side or the other of the cranium, but not exclusively, so that some bouts are on the other side.
The heritable nature of classic migraine is apparent from its occurrence in several members of the family of the same and successive generations in 60 to 80 percent of cases; the familial frequency of common migraine is slightly lower. Twin and sibling studies have not revealed a consistent mendelian pattern in either the classic or common form. Certain special forms of migraine, such as familial hemiplegic migraine, appear to be monogenic disorders but the role of these genes, most of which code for ion channels, in classic and common migraine is speculative.
Migraine, with or without aura, is a remarkably common condition. A study by Stewart and colleagues in the United States showed differences in the prevalence of migraine between individuals of white, African, and Asian origin of approximately 20, 16, and 9 percent, respectively, among women, and 9, 7, and 4 percent for men (see also Lipton et al). One-third of migraineurs have more than three attacks monthly if untreated and many require bed rest or severe curtailment of daily activities. Migraine may have its onset in childhood but usually begins in adolescence or young adulthood; in more than 80 percent of patients, the onset is before 30 years of age, and the physician should be cautious in attributing headaches that appear for the first time after this age to migraine, although there are exceptions.
In younger women, the headaches often tend to occur during the premenstrual period; in approximately 15 percent of such migraineurs, the attacks are exclusively perimenstrual (also termed "catamenial migraine"). Menstrual migraine, discussed further on, had been considered to be solely related to the withdrawal of estradiol (based on the work of Somerville). It is now acknowledged that the influence of sex hormones on headache is more complex. Migraine tends to cease during the second and third trimesters of pregnancy in 75 to 80 percent of women, and in others they continue at a reduced frequency; less often, attacks of migraine or the associated neurologic symptoms first appear during pregnancy, usually in the first trimester.
Although migraine commonly diminishes in severity and frequency with age, it may actually worsen in some postmenopausal women, and estrogen therapy may either increase or, paradoxically, diminish the incidence of headaches. The use of birth control pills is associated with an increased frequency and severity of migraine and in rare instances has resulted in a permanent neurologic deficit (see further on and Chap 34).
Some patients link their attacks to certain dietary items—particularly chocolate, cheese, fatty foods, oranges, tomatoes, and onions—but these connections have proved invalid in controlled trials and, except in the occasional persuasive instance, they seem to us to be overrated. Some of these foods are rich in tyramine, which has been incriminated as a provocative factor in migraine. Alcohol, particularly red wine or port, regularly provokes an attack in some persons; in others, headaches are fairly consistently induced by exposure to glare or other strong sensory stimuli, sudden jarring of the head ("footballer's migraine"), or by rapid changes in barometric pressure. A common trigger is excess caffeine intake or withdrawal of caffeine.
Migraine with aura frequently has its onset soon after awakening, but it may occur at any time of day. During the preceding day or so, there may have been mild changes in mood (sometimes a surge of energy or a feeling of well-being), hunger or anorexia, drowsiness, or frequent yawning. Then, abruptly, there is a disturbance of vision consisting usually of unformed flashes of white, or silver, or, rarely, of multicolored lights (photopsia). This may be followed by an enlarging blind spot with a shimmering edge (scintillating scotoma), or formations of dazzling zigzag lines (arranged like the battlements of a castle, hence the term fortification spectra, or teichopsia). Other patients complain instead of blurred or shimmering or cloudy vision, as though they were looking through thick or smoked glass or the wavy distortions produced by heat rising from asphalt. These luminous hallucinations move slowly across the visual field for several minutes and may leave an island of visual loss in their wake (scotoma); the latter is usually homonymous (involving corresponding parts of the field of vision of each eye), pointing to its origin in the visual cortex. Patients often attribute these visual symptoms to one eye rather than to parts of both fields. Ophthalmologic abnormalities of retinal and optic nerve vessels have been described in some cases but are not typical.
Other focal neurologic symptoms, much less common than visual ones, include numbness and tingling of the lips, face, and hand (on one or both sides); slight confusion of thinking; weakness of an arm or leg; mild aphasia or dysarthria, dizziness, and uncertainty of gait or drowsiness. Only one or a few neurologic phenomena are present in any given patient and they tend to occur in more or less the same combination in each attack. If weakness or paresthetic numbness spreads from one part of the body to another, or if one neurologic symptom follows another, this occurs relatively slowly over a period of minutes (not over seconds, as in a seizure, or simultaneously in all affected parts as in a transient ischemic attack).
The visual or neurologic symptoms usually last for less than 30 min, sometimes longer. As they recede, a unilateral dull pain develops of slowly increasing intensity that progresses to a throbbing headache (usually but not always on the side of the cerebral disturbance). At the peak of the pain, within minutes to an hour, the patient may be forced to lie down and to shun light (photophobia) and noise (phonophobia). Light is irritating and may be painful to the globes, or it is perceived as overly bright (dazzle) and strong odors are disagreeable. Nausea and, less often, vomiting may occur. The headache lasts for hours and sometimes for a day or even longer and is always the most unpleasant feature of the illness. The temporal scalp vessels may be tender and the headache is worsened by strain or jarring of the body or head. Pressure on the scalp vessels or carotid artery may momentarily reduce the pain and releasing pressure accentuates it.
Between attacks, the migrainous patient is normal. In the past, it was believed that a migrainous personality existed, characterized by tenseness, rigidity of attitudes and thinking, meticulousness, and perfectionism. Further analyses, however, have not established a particular personality type in the migraineur. A relationship of migraine to epilepsy in general is also tenuous; however, the incidence of seizures is slightly higher in migrainous patients and their relatives than in the general population, and there are syndromes that encompass both disorders.
Some patients note that their attacks of migraine tend to occur during the "let-down period," after many days of hard work or tension. There is an overrepresentation of motion sickness or a vague instability of vision or accommodation, sensitivity to striped patterns, fainting, and of fleeting sensory symptoms on one side of the body in migraineurs. Moreover, as appreciated by Graham, migraine has a lifetime profile and is a familial disease that includes some or many of the following: colic in infancy, motion sickness, episodic abdominal pain, fainting, alcohol sensitivity, exercise-induced headaches, "sinus headaches," "tension headaches," and menstrual headaches. These are fairly dependable markers of the disease, and their absence in the patient or family members should at least cause the consideration of alternative explanations for cranial pain.
Much variation occurs in migraine. The headache may be exceptionally severe and abrupt in onset ("crash migraine" or "thunderclap headache"), raising the specter of subarachnoid hemorrhage. Careful questioning in these cases sometimes reveals that the headache did not truly attain its peak rapidly but evolved over several minutes. Nonetheless, the distinction of this type of "thunderclap headache" from subarachnoid hemorrhage can be made only by examination of the CSF and imaging of the brain (see further on, under "Special Varieties of Headache").
A headache may at times precede or accompany, rather than follow, the neurologic abnormalities of migraine with aura. Although typically hemicranial (the French word migraine is said to be derived from megrim, which, in turn, is from the Latin hemicrania, and its corrupted forms hemigranea and migranea), the pain may be frontal, temporal, or, quite often, generalized.
Any two of the three principal components—neurologic abnormality, headache, and gastrointestinal upset—may be absent. With advancing age, for example, there is a tendency for the headache and nausea to become less severe, finally leaving only the neurologic abnormality, which itself recurs with decreasing frequency. This is also subject to great variation. One common configuration is a full-blown visual aura without subsequent headache (migraine without headache, or migraine dissocié).
Visual disturbances differ in detail from patient to patient; numbness and tingling of the lips and the fingers of one hand are probably next in frequency, followed by transient dysphasia or a thickness of speech and hemiparesis as mentioned earlier. Rarely, there is sudden but transient blindness or a hemianopia at the onset of a migraine attack, accompanied by only a mild headache.
A less-common form of the migraine syndrome with prominent brainstem symptoms was described by Bickerstaff. The patients, usually children with a family history of migraine, first develop visual phenomena like those of typical migraine except that they occupy much or the whole of both visual fields (temporary cortical blindness may occur). There may be associated vertigo, staggering, incoordination of the limbs, dysarthria, and tingling in both hands and feet, and sometimes around both sides of the mouth. These symptoms last 10 to 30 min and are followed by headache, which is usually occipital. Some patients, at the stage when the headache would have been likely to begin, may faint, and others become confused or stuporous, a state that may persist for several hours or longer. Exceptionally, there is an alarming period of coma or quadriplegia. The symptoms closely resemble those caused by ischemia in the territory of the basilar-posterior cerebral arteries—hence the name basilar artery or vertebrobasilar migraine. Subsequent studies have indicated that basilar migraine, although more common in children and adolescents, affects men and women more or less equally over a wide age range, and that the condition is not always benign and transient.
Ophthalmoplegic and Retinal Migraine
The ophthalmoplegic migraines are recurrent unilateral headaches associated with weakness of extraocular muscles. A transient third-nerve palsy with ptosis, with or without involvement of the pupil, is the usual picture; rarely, the sixth nerve is affected. This disorder is more common in children. The ocular paresis often outlasts the headache by days or weeks; after many attacks, a slight mydriasis and, rarely, ophthalmoparesis may remain permanently.
In some cases of uniocular visual disturbance with scotoma, the retinal arterioles have been reported to be attenuated and, rarely, there are retinal hemorrhages as described by Berger and colleagues. More often, there are no funduscopic changes. Such events are referred to as retinal migraine, or, more accurately, ocular migraine, as either the retinal or the ciliary circulation may be involved. However, in adults the syndrome of headache, unilateral ophthalmoparesis, and loss of vision may have more serious causes, including temporal (cranial) arteritis.
Migraine Following Head Injury
Cranial trauma of almost any degree may precipitate a migraine headache in persons prone to the condition. A particularly troublesome migraine variant occurs in a child or adolescent who, after a trivial or mild head injury, may lose vision, suffer severe headache or be plunged into a state of confusion, with belligerent and irrational behavior that lasts for hours or several days before clearing. In yet another variant, there is an abrupt onset of either one-sided paralysis or aphasia after virtually every minor head injury (we have seen this condition several times in college athletes) but without visual symptoms and little or no headache. Although a family history of migraine is frequent in such cases, there has been no history of hemiplegia in other family members.
Migraine in Young Children
This may present special difficulties in diagnosis, as a young child's capacity for accurate description is limited. Instead of complaining of headache, the child appears limp and pale and complains of abdominal pain; vomiting is more frequent than in the adult, and there may be slight fever. Recurrent attacks were referred to in the past by pediatricians as the "periodic syndrome." Another variant in the child is episodic vertigo and staggering (paroxysmal disequilibrium) followed by headache, probably a type of basilar migraine (see Watson and Steele). Also, there are puzzling patients with bouts of fever or transient disturbances in mood ("psychic equivalents") and abdominal pain (abdominal migraine), that had been attributed to migraine but are dubious entities at best. We have seen several infants and young children who have had attacks of hemiplegia (without headache), first on one side then the other, every few weeks. Recovery was complete, and arteriography in one child, after more than 70 attacks, was normal. Alternating hemiplegia of childhood may terminate in a dystonic state. The relationship of this condition to familial hemiplegic migraine (see below) remains uncertain. The only advantage of considering such attacks as migrainous is that it may protect some patients from unnecessary diagnostic procedures and surgical intervention; but, by the same token, it may delay appropriate investigation and treatment.
Familial Hemiplegic Migraine
In a related disorder, known as hemiplegic migraine, a condition mostly of infants and children (rarely adults), there are episodes of unilateral paralysis that may long outlast the headache. Several families have been described in which this condition is the result of a mutation in an ion channel (familial hemiplegic migraine; alternating hemiplegia of childhood). Of the known loci, which together account for approximately 50 percent of cases, the most common one is in the gene coding for the P/Q-type calcium channel α subunit (CACNA1A). A second locus is in the gene for the Na+/K+-adenosine triphosphatase (ATPase) channel and a rarer subtype is caused by mutations in a sodium channel α-subunit gene, SCNA1. These do not account for all cases, indicating that there are other mutations that will inevitably be discovered. It is reasonable to surmise that many of the nonfamilial cases of hemiplegic migraine are also caused by these mutations. By their nature, these channelopathies would be expected to have clinical and genetic overlap with other neurologic diseases. Indeed, there are shared traits between some of the genetic forms of familial hemiplegic migraine and both episodic and degenerative cerebellar diseases (Goadsby, 2007). Ducros and colleagues have found a variety of other neurologic features in these families, including persistent cerebellar ataxia and nystagmus in 20 percent; others had attacks of coma and hemiplegia from which they recovered.
Complicating the situation is the undoubted existence of sporadic migraine with transient hemiplegia that has no familial trait. Neurologic symptoms lasting more than an hour or so should prompt investigation for alternative causes, but none may be found. Instances of hemiplegic migraine may account for some of the inexplicable strokes in young women and older adults of both sexes, as discussed below.
Transient Ischemic Attacks and Stroke with Migraine (See Also Chap. 34.)
Attacks of migraine, instead of beginning in childhood, may have their onset later in life, and Fisher provided support for the hypothesis that some of the transient aphasic, hemianesthetic, or hemiplegic attacks of later life may be of migrainous origin ("transient migrainous accompaniment").
Rarely, migrainous neurologic symptoms, instead of being transitory, leave a prolonged or even permanent deficit (e.g., homonymous hemianopia), indicative of an ischemic stroke. This has been called complicated migraine and a small number of these prove to be migrainous infarctions. Platelet aggregation, edema of the arterial wall, increased coagulability, dehydration from vomiting, and intense, prolonged spasms of vessels have all been implicated (on rather uncertain grounds) in the pathogenesis of arterial occlusion and strokes that complicate migraine (Rascol et al).
The reported incidence of this complication has varied. At the Mayo Clinic, in a group of 4,874 patients ages 50 years or younger with a diagnosis of migraine, migraine equivalent, or vascular headache, 20 patients had migraine-associated infarctions (Broderick and Swanson). Caplan described 7 patients in whom attacks of migraine were complicated by strokes in the vertebrobasilar territory. A more recent study by Wolf and colleagues collected 17 instances of stroke and migraine. Most had a prolonged aura, either visual, sensory or aphasic and over two-thirds of the strokes, demonstrated by diffusion restriction on MRI, were in the posterior circulation territory and occurred in younger women. There is, nonetheless, a paucity of useful pathology by which to interpret the mechanism of migraine-associated stroke. The uncertain but potential role of antimigraine medications in producing stroke is discussed further on in the section on treatment. Estrogen medications have also been implicated in stroke in some women migraineurs.
In children and young adults with the mitochondrial disease MELAS (mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes) and in adults with the rare cerebral vasculopathy CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), migraine may be a prominent feature. Chap. 34 addresses these issues further.
The special problem of focal cerebral disorders associated with segmental or diffuse vasospasm, including the form that follows treatment with the "triptan" (serotonin agonist) drugs and Call-Fleming syndrome, is discussed further on in the section on treatment and under "Diffuse and Focal Cerebral Vasospasm" in Chap. 34.
A separate set of observations, mainly epidemiologic, pertain to the risk of mundane strokes in women with both migraine and cardiovascular disease later in life, and the related issue of imaging changes in migraineurs that are suggestive of small ischemic lesions. Regarding the last problem, a number of cross-sectional population studies, such as the ones by Kurth and colleagues, Scher et al, and Kruit and coworkers, indicated that MRI changes in both the deep and subcortical white matter were more frequent in women migraine patients who experienced auras than in those without auras and in the general population. A high frequency of migraine headaches was also associated in some studies with an increased number of white matter lesions. Some series have emphasized lesions in the cerebellar white matter.
In contrast, a meta-analysis of case control and cohort studies conducted by Schurks and colleagues were unable to demonstrate an increased risk for cardiovascular events. Other investigators, again depending on various population databases and few patient level
studies, have come to the opposite conclusion (Bigal et al) and suggested that all cause mortality is increased in migraine patients (Gudmundsson et al).
The implications of the ubiquitous small white matter lesions in MRI that are now familiar to neurologists are unclear. Several studies indicate that migraineurs with these changes have no greater cognitive decline over time than those in the general population. The lesions are a frequent cause for neurological consultation, sometimes with the question of multiple sclerosis having been raised. We tend to underemphasize these lesions and the risk of stroke in discussion with patients but point out that the usual stroke risk factors, smoking, hypertension, hyperlipidemia, and cardiac rhythm abnormalities should be attended to assiduously.
The issue of oral contraceptives as a risk for stroke is a more complicated matter that has not been resolved. All that can be said at the moment is that this factor did not appear to be consistent in the above discussed epidemiologic surveys, and it is the population of young women who are likely to have both exposures. The pills are not interdicted in migraineurs but perhaps lower estrogen compounds are advisable as formulations with high estrogen concentrations have been associated with clotting in the venous circulation.
Patent Foramen Ovale and Migraine
Finally, there has long been discussion of an association between migraine and patent foramen ovale. A few physicians continue to favor a causal role and have advocated closure of the foramen in an attempt to alleviate migraine. Migraine with aura has been especially associated with an open foramen. However, the largest cross-sectional (Rundek et al) and case-control (Garg et al) studies have not affirmed these associations and the issue, while still under discussion, has been of waning interest.
In some individuals, migraine attacks, for unaccountable reasons, may increase in frequency for several months. As many as three or four attacks may occur each week, leaving the scalp on one side continuously tender. An even more difficult clinical problem is posed by migraine that lapses into a condition of daily or virtually severe continuous headache (status migrainosus). The pain is initially unilateral, later more generalized, more or less throbbing, but with a constant superimposed ache and is disabling; vomiting or nausea is common at the outset but abates. Almost without exception, there is a preceding history compatible with migraine; in fact, the absence of prior headaches should raise concern about a more serious cause. Status migrainosus sometimes follows a head injury or a viral infection, but most cases have no explanation. Relief is sought by increasing the intake of ergot or serotonin agonist preparations or even opiates, often to an alarming degree, but with only temporary relief, serving at times to perpetuate the condition through a rebound mechanism.
In the diagnosis of such cases, the possibility should be considered that migraine has been combined with tension headache (migraine-tension or mixed-pattern headache) or transformed to so-called analgesia-rebound headache, or ergotamine, or serotonin agonist-dependency headache, as described by Taimi and colleagues. Narcotic addiction is another consideration. Although not generally popular, it is our practice to admit such patients to the hospital, discontinue all narcotic medications, and administer intravenous hydration, corticosteroids, one of the serotonin agonist medications, or dihydroergotamine intravenous infusion in selected patients (see further on for details of treatment). When admission is not possible or practical, the same therapeutic plan may be pursued in an ambulatory infusion center.
Migraine with CSF Pleocytosis (HaNDL)
An intriguing problem arises in the patient with migraine who is found to have a lymphocytic pleocytosis in the spinal fluid. Most of these cases in our experience have turned out to be simply instances of aseptic meningitis that have precipitated migraine in susceptible individuals. In others, a few cells are found in the spinal fluid during an attack of migraine without obvious explanation; probably a minor cellular reaction of 3 to 10 white blood cells (WBCs)/mL can be ignored if there is no fever or meningismus.
A more extensive syndrome was originally described by Bartleson, Swanson, and Whisnant under the title "A migrainous syndrome with cerebrospinal fluid pleocytosis". A series reported by Gomez-Aranda and colleagues gave the syndrome the name, "Pseudomigraine with Temporary Neurological Symptoms and Lymphocytic Pleocytosis", also called HaNDL (Headache with Neurological Deficits and CSF Lymphocytosis) by Berg and Williams, together describing what is probably yet another migraine variant. Gomez-Aranda's series comprised 50 adolescents and young adults, predominantly males, who developed several separate episodes of transient neurologic deficits lasting hours, accompanied by migraine-like headaches, sometimes with slight fever but no stiff neck. One-quarter of this group had a history of past migraine and a similar number had a viral-like illness within 3 weeks of the neurologic problem. The CSF contained from 10 to 760 lymphocytes per cubic millimeter, and the total protein was elevated. The transient neurologic deficits were mainly sensorimotor and aphasic; only 6 patients had visual symptoms. The patients were asymptomatic between attacks and in none did the entire illness persist beyond 7 weeks.
The causation and pathophysiology of this syndrome and its relation to migraine are obscure. We have observed several cases, all in otherwise healthy middle-aged men, none related to the use of nonsteroidal medications, which may cause an aseptic meningeal reaction, and we found corticosteroids to be helpful. The distinction between this syndrome and the recurrent aseptic meningitis of Mollaret and other chronic meningitic syndromes as well as cerebral vasospasm or vasculitis is difficult (see "Chronic Persistent and Recurrent Meningitis" in Chap. 32).
Cause and Pathogenesis of Migraine
So far, it has not been possible to determine from the many clinical observations and investigations, a unifying theory as to the cause and pathogenesis of migraine. Tension and other emotional states, which are claimed by some migraineurs to precede their attacks, are so inconsistent as to be no more than potential aggravating factors. Clearly, an underlying genetic factor is implicated, although it is expressed in a recognizable mendelian pattern in only a small number of families (see above). The puzzle is how this genetic predisposition is translated periodically into a regional neurologic deficit, unilateral headache, or both. For many years, our thinking about the pathogenesis of migraine was dominated by the views of Harold Wolff and others—that the headache was caused by the distention and excessive pulsation of branches of the external carotid artery. Certainly, the throbbing, pulsating quality of the headache and its relief by compression of the common carotid artery supported this view, as did the early observation of Graham and Wolff that the headache and amplitude of pulsation of the extracranial arteries diminished after the intravenous administration of ergotamine.
The importance of vascular factors continues to be emphasized by more recent findings but not in the way envisaged by Wolff. For example, in a group of 11 patients with classic migraine, Olsen and colleagues, using the xenon inhalation method, noted a regional reduction in cerebral circulation spreading forward from the occipital region during the period when neurologic symptoms appear. They concluded that the reduction in blood flow was consistent with the cortical spreading depression syndrome described below. In a subsequent study, Woods and colleagues described a patient who, during positron emission tomography (PET), fortuitously had an attack of common migraine with blurred vision. Sophisticated measurements showed a reduction in blood flow that started in the occipital cortex and spread slowly forward on both sides, in a manner much like that of the spreading cortical depression of Leão (see below) and Cutter and colleagues, using perfusion-weighted MRI, corroborated the finding of diminished occipital cerebral blood flow during the aura. However, a study using single-photon emission computed tomography (SPECT) in 20 patients during and after attacks of migraine without aura disclosed no focal changes of cerebral blood flow; also, no changes occurred after treatment of the attacks with 6 mg of subcutaneous sumatriptan (Ferrari et al, 1995).
In reference to the extracranial vessels, Iversen and associates, by means of ultrasonography, documented a dilatation of the superior temporal artery on the side of the migraine during the headache period. The same dilatation in the middle cerebral arteries has been inferred from observations with transcranial Doppler insonation. The complication of cerebral infarction is also in keeping with a vascular hypothesis, but it involves only a tiny proportion of migraineurs. The vascular hypothesis must be regarded as uncertain, but, clearly, there is frequently a reduction in posterior cortical blood flow during an aura. What is not established is whether the blood flow changes are fundamental or simply the result in a reduction in cortical activity. The original opinion expressed by Wolff that a vascular element is responsible for the cranial pain of migraine is also unconfirmed.
The relationship between the vascular changes and evolving neurologic symptoms of migraine are noteworthy. Lashley, who plotted his own visual aura, calculated that the cortical impairment progressed at a rate of 2 to 3 mm/min over the surface of the brain. Similarly, during the aura, there is a regional reduction in blood flow, as noted above. It begins in one occipital lobe and extends forward slowly (2.2 mm/min) as a wave of "spreading oligemia" that does not respect arterial boundaries (Lauritzen and Olesen). Both of these events are intriguingly similar to the above-mentioned phenomenon of "spreading cortical depression," first observed by Leão in experimental animals. He demonstrated that a noxious stimulus applied to the rat cortex was followed by vasoconstriction and slowly spreading waves of inhibition of the electrical activity of cortical neurons, moving at a rate of approximately 3 mm/min. Lauritzen and Olesen attribute both the aura and spreading oligemia to the spreading cortical depression, and considerable work since then has corroborated this idea. These observations, however, apply only to the aura.
An alternative, but not necessarily exclusive hypothesis links the aura and the painful phase of migraine through a neural mechanism originating in the trigeminal nerve as proposed by Moskowitz. This is based on the innervation of extracranial and intracranial vessels by small unmyelinated fibers of the trigeminal nerve that subserve both pain and autonomic functions (the "trigeminovascular" complex). This model provides an explanation for migraine pain in the trigeminal ganglion. Activation of these fibers releases substance P, calcitonin gene-related peptide (CGRP), and other peptides into the vessel wall, which serves to sensitize the trigeminal system to the pulsatility of cranial vessels, and to increase their permeability, thereby promoting an inflammatory response. The small molecules released from nerve endings adjacent to the cortex would then incite spreading depression in this model. Against this hypothesis is the occurrence of headache as often as not on the side opposite the side of generation of the aura and the lack of clinical effect of drugs that work in this experimental model. Most likely, both neural and vascular mechanisms are operative and they interact.
In part to address the action of the serotonin agonist drugs on migraine (see below), a body of evidence has been assembled that serotonin (5-HT) acts as a humoral mediator in the neural and vascular components of migraine headache. Serotonin is discharged from platelets at the onset of headache and the headache is reduced by the injection of 5-HT. This led to the development by Humphrey of sumatriptan, which acted selectively on 5-HT1B/D receptors so as to reduce side effects. This was the forerunner of the large group of "triptans." More recently, nitric oxide generated by endothelial cells has been implicated as the cause of the pain of migraine headache, but the reason for its release and the relationship to changes in blood flow is unclear.
Blau and Dexter and Drummond and Lance are confident that the presence or absence of headache does not depend solely on extracranial vascular factors. These authors point to their findings that occlusion of blood flow through the scalp or common carotid circulation fails to alleviate the pain of migraine in one-third to one-half the patients. Lance (1998) has suggested that the trigeminal pathways are in a state of persistent hyperexcitability in the migraine patient and that they discharge periodically, perhaps in response to hypothalamic stimulus acting on the endogenous pain control pathways. This is in keeping with current theories regarding the trigeminovascular complex discussed above, as well as with evolving ideas on central sensitization to pain because of repeated noxious stimulation from one body region that may produce a type of centrally mediated allodynia. The role of alternative factors in migraine has been reviewed by Lance and Goadsby and in a pictorial representation by Goadsby and colleagues.
The foregoing observations leave a number of questions unanswered. Is one to conclude that migraine with and without aura are different diseases, involving extracranial arteries in one instance and intracranial ones in another? Is the circulatory change the primary cause of headache, or is it a secondary or coincidental phenomenon? Is diminished neuronal activity (spreading depression) the primary cause of neurologic symptoms (it seems so) and headache (unclear), and is the diminished regional blood flow secondary to reduced metabolic demand? Why are the posterior portions of the brain (visual auras) so often implicated (perhaps because of richer trigeminal innervation of the posterior vessels)? The neural mechanisms that underlie these changes and precisely what is altered by the genetic predisposition to migraine are unresolved. No final reconciliation of all these data is possible and migraine remains incompletely explained.
Migraine with aura should occasion no difficulty in diagnosis if a proper history is obtained. Most often, the symptoms begin as "positive," i.e., scintillation, paresthesia, as opposed to the later "negative" scotoma, numbness, aphasia, or paresis. The difficulties come from a lack of awareness that a progressively unfolding neurologic syndrome may be migrainous in origin and may occur without headache. Furthermore, recurrent migraine headaches take many forms, some of which may prove difficult to distinguish from the other common types of headache, and it is not generally recognized that migraine headaches need not be severe or disabling. Some of these problems merit elaboration because of their practical importance.
The neurologic part of the migraine syndrome may resemble a transient ischemic attack, focal epilepsy, the clinical effects of a slowly evolving hemorrhage from an arteriovenous malformation, or a thrombotic or embolic stroke. It is the pace of the neurologic symptoms of migraine that distinguish it from epilepsy and most cases of stroke. Furthermore, the positive rather than ablative nature of the symptoms assists in distinguishing it from the usual stroke syndromes.
Ophthalmoplegic migraine may suggest a carotid-cavernous or supraclinoid aneurysm. Transient monocular blindness from carotid stenosis is infrequent in the age group affected most by migraine, but the antiphospholipid syndrome, which has some ill-defined relationship to migraine, does cause episodic unilateral visual loss in this group and should be sought as the explanation for transient monocular blindness with or without headache. The invariant occurrence of migraine on the same side of the head increases the likelihood of an underlying arteriovenous malformation (AVM) or other structural lesion. R.D. Adams, who studied more than 1,200 patients with AVM, also found that the headaches, which occurred in more than 30 percent of these individuals, usually did not include the other features of either migraine or cluster headache. However, in about 5 percent, the headaches were associated with visual aura, making them indistinguishable from neurologic migraine. In most, the AVM was in the occipital region and on the side of the headache. Approximately half of the patients with AVM and migraine had a family history of migraine. It is unclear to us if AVM can be regarded as an acknowledged cause of recurrent migraine-like headache. It is, of course, possible that given the ubiquity of migraine in the population, that the association is coincidental.
This topic may be divided into two parts—control of the individual acute attack, and prevention that includes both medications and lifestyle modifications. The time to initiate treatment of an attack is during the neurologic (visual) prodrome or at the very onset of the headache (see below). If the headaches are mild, the patient may already have learned that aspirin, acetaminophen, or another nonsteroidal anti-inflammatory drug (NSAID) will suffice to control the pain. Insofar as a good response may be obtained from one type of NSAID and not another, it may be advisable to try two or three preparations in several successive attacks of headache and to use moderately high doses if necessary. Some of our colleagues state that reliable patients can be given small amounts of codeine or oxycodone, usually combined with aspirin or acetaminophen, for limited periods. The combination of aspirin or acetaminophen, caffeine, and butalbital, although popular with some patients, is usually incompletely effective if the headache is severe and is also capable of causing drug dependence. Numerous other medications have proved effective and each has had a period of popularity among neurologists and patients.
For severe attacks of migraine headache, sumatriptan or one of the other serotonin agonist "triptans" in this class (e.g., zolmitriptan, rizatriptan, naratriptan, almotriptan, eletriptan, and frovatriptan), or the ergot alkaloids, ergotamine tartrate, and particularly dihydroergotamine (DHE), are the most effective forms of treatment and are best administered early in the attack, ideally just after an aura or at the onset of headache.
Patients with waning visual auras should be advised to wait to self-administer subcutaneous serotonin agonists until the headache begins. Clinical experience and the study by Bates and colleagues suggests that the triptans are ineffective in preventing headache if given during the aura; they are albeit, safe (see below). In contrast, the slower acting nasal spray or oral formulations are often ineffective if given too long after the start of headache and patients have learned to administer them during the aura and, again, this seems safe.
A single 6-mg dose of sumatriptan or its equivalent, given subcutaneously, is an effective and well-tolerated treatment for migraine attacks. When successful, it eliminates or reduces the accompanying symptoms of nausea, vomiting, photophobia, and phonophobia. An advantage of the serotonin agonist drugs, aside from their relative safety, is the ease of self-administration using prepackaged injection kits, thus avoiding frequent and inconvenient visits to the emergency department. For example, sumatriptan can also be given orally in a 25- or 50-mg tablet and as a nasal spray (20 mg per spray), zolmitriptan in a 2.5- or 5-mg tablet, and rizatriptan in a 5- or 10-mg dose tablet, repeated, if needed, in 2 h. For these oral preparations, latency for headache relief is longer than with subcutaneous injection or inhalation. If one serotonin agonist is found to be ineffective, another drug or an alternative route of administration, such as intranasal, may be tried. These medications are summarized in Table 10-2.
Table 10-2 Triptans for Oral Use ||Download (.pdf)
A large and often cited meta-analysis of the available drugs in 53 separate trials conducted by Ferrari and colleagues (2001) found modest differences in overall efficacy between drugs. Loder has given a tabulated comparison of the main drugs for migraine and a review of their use in routine situations. Others in this class are sure to be developed in the future and subtle, but usually clinically minor, differences between them undoubtedly will be highlighted. Sumatriptan is available as a nasal spray, which is useful in patients with nausea and vomiting. The response rate after 2 h is similar to that of the orally administered drug, and the nasal spray acts more rapidly.
Ergotamine is an equally effective agent, but its peripheral and coronary vasoconstricting side effects have reduced its use. This is an alpha-adrenergic agonist with strong serotonin receptor affinity and vasoconstrictive action. The drug is taken as an uncoated 1- to 2-mg tablet of ergotamine tartrate, held under the tongue until dissolved (or swallowed), or in combination with caffeine. Repeat use is not advisable as it may lead to prolonged or daily headache. A single oral dose of promethazine 50 mg, or of metoclopramide 20 mg, given with the ergotamine, relaxes the patient and allays the potential nausea and vomiting from ergotamine. Patients in whom vomiting prevents oral administration may be given ergotamine by rectal suppository or DHE by nasal spray or inhaler (one puff at onset and another at 30 min) or can learn to give themselves a subcutaneous injection of DHE (usual dosage, 1 mg). Caffeine, 100 mg, is thought, on slim evidence, to potentiate the effects of ergotamine and other medications for migraine. When ergotamine is administered early in the attack, the headache will be abolished or reduced in severity and duration in 70 to 75 percent of patients.
An important problem pertains to the magnitude of risk of stroke from serotonin agonists in patients with prolonged visual aura or other neurologic symptoms that persist during the period of headache, or focal neurologic symptoms that are possibly attributable to migraine because of a past history of migraine with or without aura. As a matter of course, serotonin agonists and ergots are generally avoided if there is an ongoing and prolonged aura of any type, including visual, but particularly with hemiparesis, aphasia, or features such as vertigo, drowsiness, or diplopia, referable to the basilar artery. Not all experts agree with this proscription and some small series, among them 13 patients reported by Klapper and colleagues, have found triptans safe to use if a headache with neurologic signs has commenced, but this issue has not been resolved. As previously noted, although this class of drugs may not be helpful during the visual aura, they also seem to do no harm (see Bates et al).
Ergot drugs and triptans are contraindicated in symptomatic and asymptomatic coronary artery disease and poorly controlled hypertension. Rare cases of severe but reversible cerebral vasospasm have been reported after the use of ergotamine or a serotonin agonist drug, but most of these patients in fact had not had neurologic features as part of their initial headache syndrome. Of particular danger, however, is the often unnoticed, concurrent use of other sympathomimetic drugs such as phenylpropanolamine as in one of the cases described by Singhal and colleagues and by Meschia and associates (see discussion of Call-Fleming syndrome, "Diffuse Vasoconstriction," "Diffuse and Focal Cerebral Vasospasm" in Chap. 34). Cerebral hemorrhage is another rare complication of serotonin agonist use that possibly relates to hypertension induced by triptans or ergots.
For severely ill patients who arrive in the emergency department or physician's office, having failed to obtain relief from a prolonged headache with the above medications, Raskin (1986) has found metoclopramide 10 mg IV, followed by DHE 0.5 to 1 mg IV every 8 h for 2 days, to be effective. We also use this approach in cases of status migrainosus. The administration of intravenous DHE can be combined with a lidocaine infusion. The potential success of metoclopramide alone should not be dismissed, as we and others have occasionally found that the headache abates after this initial injection. The sympathomimetic drug isometheptene combined with a sedative, and acetaminophen (Midrin) has been useful for some patients and probably acts in a similar way to ergotamine and sumatriptan. A wide array of other drugs including almost all of the conventional nonsteroidal anti-inflammatory medications has been recommended as adjunctive therapy, e.g., prochlorperazine, chlorpromazine, ketorolac, and intranasal lidocaine. Each of these drugs, given alone, is effective in alleviating the headache in about half of patients, emphasizing the need for blinded placebo-controlled trials for any new drug that is introduced for the treatment of headache.
Intravenous and oral corticosteroids have been found anecdotally to be useful in refractory cases and as a means of terminating migraine status, but they should not be given continuously. In a randomized trial of intravenous dexamethasone 10 mg in an emergency department setting, Friedman et al found no benefit. As an alternative to steroids and more commonly used non-steroidal agents, Weatherall and colleagues used intravenous aspirin (lysine acetylsalicylate, 1 g, repeated up to five times) with reasonably good effect in inpatient management of migraine and other headache disorders. We have determined that this agent is difficult to obtain from hospital pharmacies.
If, in an individual attack, all of the foregoing measures fail, it is probably best to resort briefly to narcotics, which usually give the patient a restful, pain-free sleep. Halfway measures at this point are usually futile. However, the use of narcotics as the mainstay of acute or prophylactic therapy is to be avoided. As mentioned above, if the pain does not abate in 12 to 24 h, corticosteroids in any of several regimens may be added and continued for several days.
Based on the action of certain peptides in the trigeminovascular complex, novel antagonists of CGRP have been used with some success (Olesen et al). Drugs of this type as well as inducible nitric oxide synthase (iNOS) inhibitors and receptor blockers that work by a different mechanism than do the serotonin agonists may be alternatives in the future. So far, these have not been effective as prophylactic drugs but are currently being studied for acute treatment of migraine.
In individuals with frequent migrainous attacks, efforts at prevention are worthwhile. The survey by Lipton and colleagues, found approximately one-fourth of patients were appropriate for some form of prophylactic treatment on the basis of the frequency and severity of their headaches, usually more than one severe episode per week. The most effective agents have been beta-adrenergic blockers, certain antiepileptic drugs, and tricyclic antidepressants. Often, comorbidities such as depression, hypertension, epilepsy or coronary artery disease guide the choice among these three classes of drugs. Some headache specialists have expressed the opinion that amitryptiline may be more effective than the others if headaches are very frequent and that propranolol is more effective if severity of headaches is the prime concern.
Considerable success has been obtained with propranolol, beginning with 10 to 20 mg two to three times daily and increasing the dosage gradually to as much as 240 mg daily, probably best given as a long-acting preparation in the higher dosage ranges. Under-dosing is a major reason for ineffectiveness. If propranolol is unsuccessful or not tolerated, one of the other beta blockers, specifically those that lack agonist properties—atenolol (40 to 160 mg/d), timolol (20 to 40 mg/d), or metoprolol (100 to 200 mg/d)—may prove to be effective. In patients who do not respond to these drugs over a period of 4 to 6 weeks, valproic acid 250 mg taken three to four times daily, other antiepileptic drug such as topiramate, or amitriptyline, 25 to 125 mg nightly may be tried. The newer antidepressants (e.g., specific serotonin reuptake inhibitors) are not as effective and may even cause headache in our experience. Calcium channel blockers (e.g., verapamil, 320 to 480 mg/d; nifedipine, 90 to 360 mg/d) are also reportedly effective in decreasing the frequency and severity of migraine attacks in some patients, but there is typically a lag of several weeks before benefit is attained and our success with them has been limited. Isometheptene (Midrin) as already mentioned; indomethacin, 150 to 200 mg/d; and cyproheptadine (Periactin), 4 to 16 mg/nightly are found to be helpful in some patients and may be particularly useful in preventing predictable attacks of perimenstrual migraine. A typical experience is for one of these medications to reduce the number and severity of headaches for several months and then to become less effective, whereupon an increase in the dosage, if tolerated, may help; or one of the many alternatives can be tried. The newest putative treatment for chronic or frequently repeating headaches, both migraine and tension, is the injection of botulinum toxin (Botox) into sensitive temporalis and other cranial muscles. Elimination of headaches for 2 to 4 months has been reported—a claim that justifies further study. Whether these injections are of value during an acute attack is uncertain. Surgical decompression of sensory nerves in the scalp and related techniques have also been advocated but require rigorous study.
Methysergide (Sansert), an ergot derivative that was more widely used in the past, in doses of 2 to 6 mg daily for several weeks or months is effective in the prevention of migraine. Retroperitoneal and pulmonary fibrosis are rare but serious complications that can be avoided by discontinuing the medication for 3 to 4 weeks after every 5-month course of treatment. Its use has virtually ceased.
Some patients allege that certain items of food induce attacks (chocolate, peanuts, hot dogs, smoked meats, oranges, and red wine are the ones most commonly mentioned), and it is obvious enough that they should avoid these foods if possible. Limiting caffeinated beverages may be helpful. In certain cases, the correction of a refractive error, an elimination diet, or behavioral modification is said to have reduced the frequency and severity of migraine and of tension headaches. However, the methods of study and the results have been so poorly controlled that it is difficult to evaluate them. All experienced physicians appreciate the importance of helping patients rearrange their schedules with a view to controlling tensions and hard-driving lifestyles. There is no single program to accomplish this. Psychotherapy has not been helpful, or at least one can say that there is no evidence of its value. The claims for sustained improvement of migraine with chiropractic manipulation are similarly unsubstantiated and do not accord with our experience. Meditation, acupuncture, and biofeedback techniques all have their advocates, but again, the results, while not to be entirely discounted, are uninterpretable.
This is a group of relatively uncommon syndromes that may be allied with migraine but respond very well and specifically to indomethacin both acutely and as prophylaxis, so much so that some authors have defined a category of indomethacin-responsive headaches. These include orgasmic migraine, chronic paroxysmal hemicrania (see further on), hemicrania continua, exertional headache, hypnic headache, brief head pains (jabs and jolts and "ice-pick" headaches), and some instances of premenstrual migraine. These are summarized in Table 10-3.
Table 10-3 Indomethacin Responsive Headaches ||Download (.pdf)
Table 10-3 Indomethacin Responsive Headaches
Valsalva related headaches
Primary headache associated with sexual activity*
Primary exertional (exercise induced, weight lifters) headache*
Primary cough headache
Chronic paroxysmal hemicrania
Episodic paroxysmal hemicrania
SUNCT (short-lasting unilateral neuralgiform attacks with conjunctival injection and tearing)
"Jabs and jolts"
Idiopathic stabbing (ice pick) headache
This type of headache has been described in the past under a variety of names, including paroxysmal nocturnal cephalalgia, migrainous neuralgia, histamine cephalalgia (Horton's headache), and others. Kunkle and colleagues, who were impressed with the characteristic temporal "cluster pattern" of the attacks, coined the term in current use—cluster headache. This headache pattern occurs predominantly in adult men (age range: 20 to 50 years; male-to-female ratio approximately 5:1) and is characterized by a severe consistent unilateral orbital localization. The pain is felt deep in and around the eye, is very intense and nonthrobbing as a rule, and often radiates into the forehead, temple, and cheek—less often to the ear, occiput, and neck. Its denominative feature is the nightly recurrence, between 1 and 2 h after the onset of sleep, or several times during the night for several or more consecutive days; thus "cluster". Less often, it occurs during the day or early evening, unattended by aura or vomiting. The headache has been called the "alarm clock headache" because it may recur with remarkable regularity each night for periods extending as long as many weeks, followed thereafter by complete freedom for many months or even years. However, in approximately 10 percent of patients, the headache becomes chronic, persisting over days, months, or even years.
There are several associated vasomotor phenomena by which cluster headache can be identified: a blocked nostril, rhinorrhea, injected conjunctivum, lacrimation, miosis, and a flush and edema of the cheek, all lasting on average for 45 min (range: 15 to 180 min). Some of our patients, when alerted to the sign, also report a slight ptosis on the side of the orbital pain; in a few, the ptosis has become permanent after repeated attacks. The homolateral temporal artery may become prominent and tender during an attack, and the skin over the scalp and face may be hyperalgesic.
Most patients arise from bed during an attack and sit in a chair and rock or pace the floor, holding a hand to the side of the head. The pain of a given attack may leave as rapidly as it began or may fade away gradually. Almost always the same orbit is involved during a cluster of headaches as well as in recurring bouts. During the period of freedom from pain, alcohol, which commonly precipitates headaches during a cluster, no longer has the capacity to do so. The picture of cluster headache, including the patient's nocturnal behavior in response to it, is usually so characteristic that it cannot be confused with any other disease, although those unfamiliar with it may entertain a diagnosis of migraine, trigeminal neuralgia, carotid aneurysm, or temporal arteritis.
A somewhat similar syndrome is produced by the Tolosa-Hunt syndrome of eye pain and ocular motor paralysis caused by dural granuloma at the orbital apex (see further on) and the paratrigeminal syndrome of Raeder, which consists of paroxysms of pain somewhat like that of tic douloureux in the distribution of the ophthalmic and maxillary divisions of the fifth nerve, in association with unilateral Horner syndrome (ptosis and miosis but with preservation of facial sweating). Loss of sensation in a trigeminal nerve distribution and mild weakness of muscles innervated by the fifth nerve are often added. Raeder syndrome is now recognized as a heterogeneous syndrome, some cases being cluster and others caused by a structural lesion in or near the carotid siphon.
Trigeminal Autonomic Cephalgias (Cluster Variants)
Cases of paroxysmal pain behind the eye or nose or in the upper jaw or temple—associated with blocking of the nostril or lacrimation and described in the past under the titles of sphenopalatine (Sluder), petrosal, vidian, and ciliary neuralgia (Charlin or Harris)—probably represent variants of cluster headache. A similar head pain may occasionally be confined to the lower facial, postauricular, or occipital areas. Ekbom distinguished yet another "lower cluster headache" syndrome with infraorbital radiation of the pain, an ipsilateral partial Horner syndrome, and ipsilateral hyperhidrosis. There is no evidence to support the separation of these neuralgias as distinct entities, and they have collectively been called trigeminal autonomic cephalgias. They are important, however, because of the high frequency of underlying intracranial lesions. Favier and colleagues collected 4 of their own cases and 27 from the literature to emphasize the range of underlying diseases, including intracranial aneurysms, peritentorial or parasellar meningiomas, or other tumors and nasopharyngeal cancers surrounding the carotid artery. We have encountered a case of Wegener granulomatosis of the soft palate that presented as a paroxysmal trigeminal autonomic neuralgia. The headache syndrome disappeared with cyclophosphamide treatment of the underlying granulomatous disorder.
Chronic paroxysmal hemicrania was the name given by Sjaastad and Dale to a rapidly repetitive unilateral form of headache that resembles cluster headache in many respects but has several distinctive features. These are of much shorter duration (2 to 45 min) than cluster and usually affect the temporoorbital region of one side, accompanied by conjunctival hyperemia, rhinorrhea, and in some cases a partial Horner syndrome. Even periorbital ecchymosis may accompany a severe attack. Unlike cluster headache, however, the paroxysms occur many times each day, recur daily for long periods (the patient of Price and Posner had an average of 16 attacks daily for more than 40 years), and, most important, respond dramatically to the administration of indomethacin, 25 to 50 mg tid. Unlike cluster headache, chronic paroxysmal hemicrania is more common in women than in men (ratio of 3:1).
The acronym SUNCT (short-lasting unilateral neuralgiform attacks with conjunctival injection and tearing) has been applied to an episodic condition with attacks of even briefer duration, but otherwise similar to paroxysmal hemicrania in which the supraorbital or temporal pain lasts up to 4 min or so and frequent; it does not usually respond to indomethacin.
A similar hemicrania but without autonomic features, may be symptomatic of lesions near the cavernous sinus (mainly pituitary adenoma) or in the posterior fossa, but most cases are idiopathic. The typical episode of pain lasts approximately 20 min. Also known is a recurrent nocturnal headache in elderly individuals ("hypnic headache"), as described further on.
The relationship of cluster headache and all of its variants to migraine remains conjectural. No doubt the headaches in some persons have some of the characteristics of both, hence the terms migrainous neuralgia and cluster migraine (Kudrow). Lance and others, however, have pointed out differences that seem important to us: flushing of the face on the side of a cluster headache and pallor in migraine; increased intraocular pressure in cluster headache, normal pressure in migraine; increased skin temperature over the forehead, temple, and cheek in cluster headache, decreased temperature in migraine; and notable distinctions in sex distribution, age of onset, rhythmicity, and other clinical features, but prominently by differences among them in response to specific treatments. Cluster may be triggered in sensitive patients by the use of nitroglycerin and, as mentioned, by alcohol.
The cause and mechanism of the cluster headache syndrome are unknown. Gardner and coworkers originally postulated a paroxysmal parasympathetic discharge mediated through the greater superficial petrosal nerve and sphenopalatine ganglion. These authors obtained inconsistent results by cutting the nerve, but others (Kittrelle et al) reported that application of cocaine or lidocaine to the region of the sphenopalatine fossa (via the nostril) consistently aborts attacks of cluster headache. Capsaicin, applied over the affected region of the forehead and scalp, may have the same effect. Stimulation of the ganglion is said to reproduce the syndrome. Kunkle, on the basis of a large personal experience, concluded that the pain arises from the internal carotid artery, in the canal through which it ascends in the petrous portion of the temporal bone. In the course of an arteriogram, during which a patient with cluster headaches fortuitously developed an attack, Ekbom and Greitz noted a narrowing of the artery that was interpreted as being caused by swelling of the arterial wall, which, in turn, compromised the pericarotid sympathetic plexus and caused the Horner syndrome. This remains to be confirmed.
The cyclic nature of the attacks has been linked to a hypothalamic mechanism that governs the circadian rhythm. At the onset of the headache, the region of the suprachiasmatic nucleus appears to be active on PET (May et al). Hypothalamic activation has also been found in migraine, SUNCT, chronic paroxysmal hemicrania, and hemicrania continua. Moreover, stimulation of the hypothalamus has proved effective, although highly experimental, in stopping chronic cluster headache and SUNCT (see Leone et al and Bartsch et al).
Much was made in the past of the fact that cluster headaches could be reproduced by the intravenous injection of 0.1 mg histamine, but the effect was probably nonspecific. Goadsby has reviewed the pathophysiology of the cluster headache syndrome.
Treatment of Cluster Headache
Inhalation of 100 percent oxygen via mask for 10 to 15 min at the onset of cluster headache may abort the attack, but this is not always practical. Termination of a cycle of cluster can also be achieved with verapamil, starting with 80 mg qid and increasing the dose over days, but electrocardiogram (ECG) monitoring is recommended in the older individual. The usual nocturnal attacks of cluster headache can be treated with a single anticipatory dose of ergotamine at bedtime (2 mg orally) or with possibly lesser efficacy, an equivalent dose of serotonin agonist. Intranasal lidocaine or sumatriptan (or zolmitriptan as for migraine, see above) can also be used to abort an acute attack. In other patients, ergotamine given once or twice during the day, before an attack of pain is expected, has been helpful.
With regard to prevention of cluster headache, if ergotamine and sumatriptan are ineffective or become ineffective in subsequent bouts, many headache experts prefer to use verapamil, up to 480 mg per day. Ekbom introduced lithium therapy for cluster headache (600 mg, up to 900 mg daily), and Kudrow has confirmed its efficacy in chronic cases. Lithium and verapamil may be given together, but lithium toxicity is a frequent problem. A course of prednisone, beginning with 75 mg daily for 3 days and then reducing the dose at 3-day intervals, has been beneficial in many patients. Usually, it can be decided within a week if any one of these medications is effective. In brief, no method is effective in all cases, but the best initial approach probably involves the use of one of the triptan compounds. Rare cases of intractable cluster headache, in which the syndrome persists for weeks or longer without remission, have been treated by partial section of the trigeminal nerve, as described by Jarrar and colleagues, but these ablative measures are now always a last resort, especially when hypothalamic stimulation has been shown to be possibly effective, as mentioned earlier.
This, said to be the most common variety of headache, is usually bilateral, with occipitonuchal, temporal, or frontal predominance, or diffuse extension over the top of the cranium. The pain is usually described as dull and aching, but questioning often uncovers other sensations, such as fullness, tightness, or pressure (as though the head were surrounded by a band or clamped in a vise) or a feeling that the head is swollen and may burst. On these sensations, waves of aching pain are superimposed. These may be interpreted as paroxysmal or throbbing and, if the pain is slightly more on one side, the headache may suggest a migraine without aura. However, absent in tension headache are the persistent throbbing quality, nausea, photophobia, phonophobia, and clear lateralization of migraine. Nor do most tension headaches seriously interfere with daily activities, as migraine does. The onset is more gradual than that of migraine, and the headache, once established, may persist with only mild fluctuations for days, weeks, months, or even years. In fact, this is the only type of headache that exhibits the peculiarity of being present throughout the day, day after day, for long periods of time for which the term chronic tension-type headache is used. There is often self-acknowledged anxiety and depression, as noted below. Although sleep is usually undisturbed, the headache develops soon after awakening, and the common analgesic remedies have limited effect if the pain is of more than mild to moderate severity.
The incidence of tension headache is certainly greater than that of migraine. However, most patients treat tension headaches themselves and do not seek medical advice. Like migraine, tension headaches are more common in women than in men. Unlike migraine, they infrequently begin in childhood or adolescence but are more likely to arise in middle age and to coincide with anxiety, fatigue, and depression in the trying times of life. In the large series reported by Lance and Curran, about one-third of patients with persistent tension headaches had readily recognized symptoms of depression. They carried out a controlled and blinded trial that demonstrated benefit from amitriptyline even in those patients who were not depressed. In our experience, chronic anxiety or depression of varying degrees of severity is present in the majority of patients with protracted headaches. Migraine and traumatic headaches may, of course, be complicated by tension headache, which, because of its persistence, often arouses fears of a brain tumor or other intracranial disease. However, as Patten points out, not more than one or two patients out of every thousand with tension headaches will be found to harbor an intracranial tumor, and its discovery has been most often incidental (see further on).
In a substantial group of patients with chronic daily headache, the pain, when severe, develops a pulsating quality, to which the term tension-migraine or tension-vascular headache has been applied (Lance and Curran). Observations such as these have tended to blur the sharp distinctions between migrainous and tension headaches in some cases.
For many years, it was thought that tension headaches were a result of excessive contraction of craniocervical muscles and an associated constriction of the scalp arteries. However, it is not clear that either of these mechanisms contributes to the genesis of tension headache, at least in its chronic form. In most patients with tension headache, the craniocervical muscles are quite relaxed (by palpation) and show no evidence of persistent contraction when measured by surface electromyographic (EMG) recordings. Anderson and Frank found no difference in the degree of muscle contraction between migraine and tension headache. However, using an ingenious laser device, Sakai and associates have reported that the pericranial and trapezius muscles are hardened in patients with tension headaches. Recently, nitric oxide has been implicated in the genesis of tension-type headaches, specifically by creating a central sensitization to sensory stimulation from cranial structures. The strongest support for this concept comes from several reports that an inhibitor of nitric oxide reduces muscle hardness and pain in patients with chronic tension headache (Ashina et al). At present, these are interesting but speculative ideas.
Treatment of Tension Headache
Simple analgesics, such as aspirin or acetaminophen or other NSAIDs, may be helpful, if only for brief periods. Persistent or frequent tension headaches respond best to the cautious use of one of several drugs that relieve anxiety or depression such as amitriptyline given as a single dose at night, especially when symptoms of these conditions are present. Stronger analgesic medication should be avoided. Raskin reports success with calcium channel blockers, phenelzine, and cyproheptadine. Ergotamine and propranolol are ineffective unless there are symptoms of both migraine and tension headache. Some patients respond to ancillary measures such as massage, meditation, and biofeedback techniques. Relaxation techniques may be helpful in teaching patients how to deal with underlying anxiety and stress. Gradual withdrawal of daily doses of analgesics, ergotamines, or triptan medications is an important aspect of treating chronic daily headache.
In several surveys, headache with onset in the elderly age period was found to be a prominent problem in as many as 1 of 6 persons, and more often to have serious import than headache in a younger population. In a series reported by Pascual and Berciano, more than 40 percent were classified as having tension headaches (women more than men), and there was a wide variety of diseases in the others (posttraumatic headaches, cerebrovascular disease, intracranial tumors, cranial arteritis, severe hypertension). Cough-induced headaches and cluster headaches were present in some of the men. New-onset migraine in this age group was a rarity.
Raskin described a headache syndrome in older patients that shares with cluster headache a nocturnal occurrence (hypnic headache). It also may occur with daytime naps. However, it differs in being bilateral and unaccompanied by lacrimation and rhinorrhea. He has successfully treated a number of his patients with 300 mg
of lithium carbonate or 75 mg of sustained-release indomethacin at bedtime. The nosologic position of this hypnic headache syndrome is undetermined.
Despite these considerations, the most treacherous and neglected cause of headache in the elderly is temporal (cranial) arteritis with or without polymyalgia rheumatica, as discussed further on.
Headache and Other Craniofacial Pain with Psychiatric Disease
The most common cause of generalized persistent headache, both in adolescents and adults, is probably mild depression or anxiety in one of its several forms. A small group of older patients has delusional symptoms involving pain and physical distortion of cranial structures. As the psychiatric symptoms subside, the headaches usually disappear. Odd cephalic pains, e.g., a sensation of having a nail driven into the head (clavus hystericus), may occur in hysteria or psychosis and raise perplexing problems in diagnosis. The bizarre character of these pains, their persistence in the face of every known therapy, the absence of other signs of disease, and the presence of other manifestations of psychiatric disease provide the basis for correct diagnosis. Older children and adolescents sometimes have peculiar behavioral reactions to headache: screaming, looking dazed, clutching the head with an agonized look. Usually, migraine is the underlying disorder in these cases, the additional manifestations responding to therapeutic support and suggestion.
Severe, chronic, continuous, or intermittent headaches lasting several days or weeks appear as the cardinal symptom of several posttraumatic syndromes, separable in each instance from the headache that immediately follows head injury (i.e., scalp laceration and cerebral contusion with blood in the CSF or increased intracranial pressure).
The headache of chronic subdural hematoma is deep seated, dull, steady, mainly unilateral and may be accompanied or followed by drowsiness, confusion, and fluctuating hemiparesis. In more acute subdural hematomas, we have been impressed with the positional worsening of pain in some patients after lying down or leaning the head to one side. Tentorial hematomas produce the additional feature of pain in the eye. The head injury that gives rise to a subdural hematoma may have been minor, as described in Chap. 35, and forgotten by the patient and family. Typically, the headache increases in frequency and severity over several weeks or months. Patients who have received anticoagulation are particularly at risk. Diagnosis is established by CT or MRI.
Chronic headache is certainly a prominent feature of the postconcussion syndrome, comprising dizziness, fatigability, insomnia, nervousness, irritability, and inability to concentrate (a syndrome that we have also called posttraumatic nervous instability). This type of headache and associated symptoms, which resemble the tension headache syndrome, are described fully in Chap. 35, "Craniocerebral Trauma." The International Headache Society has classified persistence in this context as headache for longer than 3 months after injury. The patient with postconcussion syndrome requires supportive therapy in the form of repeated reassurance and explanations of the benign nature of the symptoms, a program of increasing physical activity, and the use of drugs that allay anxiety and depression. The early settlement of litigation, which is often an issue, works to the patient's advantage.
Tenderness and aching pain sharply localized to the scar of a long previous scalp laceration or surgical incision represent in a different problem and raise the question of a traumatic neuralgia or neuroma. Tender scars from scalp lacerations may be treated by repeated subcutaneous injections of local anesthetics, such as 5 mL of 1 percent procaine, which also acts as a diagnostic test.
With whiplash injuries of the neck, there may be unilateral or bilateral retroauricular or occipital pain, probably as a result of stretching or tearing of ligaments and muscles at the occipitonuchal junction or of a worsening of a preexisting cervical arthropathy. Much less frequently, cervical intervertebral discs and nerve roots are involved. However, it is questionable if chronic headache and vague neuropsychiatric symptoms can be attributed to whiplash (see Maleson).
One should also be alert to headache as a sign of carotid artery dissection after head or neck injury.
It remains a popular notion that headache is a significant symptom in many patients with brain tumor, but it is actually infrequent, particularly as the heralding symptom of a tumor in an adult. While headache is sometimes stated to occur in one-third of brain tumor cases, this is certainly the result of the high frequency of cranial imaging in headache patients. Headache probably only arises if the tumor displaces major cerebral vessels or blocks the flow of CSF, but we have seen exceptions. The pain has no specific features; it tends to be deep seated, usually non-throbbing (occasionally throbbing), and is described as aching or bursting. However, a major change in the pattern of an accustomed headache syndrome should raise suspicion of a structural lesion in the cranium. Physical activity and changes in position of the head may provoke pain, whereas rest sometimes diminishes it. Nocturnal awakening because of pain occurs in only a small proportion of brain tumor patients and is by no means diagnostic. Most headaches that awaken people at night are cluster-like headaches, hypnic headaches in the elderly, or those caused by caffeine withdrawal. Unexpected forceful (projectile) vomiting may punctuate brain tumor headache in its later stages, particularly in children, or occur early if the mass is in the posterior fossa.
If unilateral, the headache is nearly always on the same side as the tumor. Pain from supratentorial tumors is felt anterior to the interauricular circumference of the skull; from posterior fossa tumors, it is felt behind this line. Bifrontal and bioccipital headaches from tumor coming on after unilateral headaches probably signify the development of increased intracranial pressure or hydrocephalus.
Having stated that headache is not to be equated with brain tumor, one cannot help but be impressed with its frequency as a manifestation of colloid cysts, and we have several times stumbled on the correct diagnosis when an odd, unexplained bilateral headache led to brain imaging. The mechanism of headache in cases of colloid cyst, if such a relationship is valid, is not simply one of blocking the flow of CSF at the foramina of Monro, as it is not predicated on the development of hydrocephalus. Additionally, Harris described headaches of paroxysmal type with intra- and periventricular brain tumors, and many others have commented on the same type of headache with parenchymal tumors. These are severe headaches that reach their peak intensity in a few seconds, last for several minutes or as long as an hour, and then subside quickly. When they are associated with vomiting, transient blindness, leg weakness causing "drop attacks," and loss of consciousness, there is a possibility of brain tumor with greatly elevated intracranial pressure. With respect to its onset, this headache almost resembles that of subarachnoid hemorrhage, but the latter is far longer-lasting and even more abrupt in onset. In its entirety, this paroxysmal headache is most typical of the aforementioned colloid cyst of the third ventricle, but it can occur with other tumors as well, including craniopharyngiomas, pinealomas, and cerebellar masses.
Headaches of Temporal Arteritis (Giant Cell Arteritis) (See Also Chap. 34)
This type of inflammatory disease of cranial arteries is an important cause of headache in older persons. All of our patients have been older than 55 years of age, most of them older than age 65. From a state of normal health, the patient develops an increasingly intense throbbing or nonthrobbing headache, often with superimposed sharp, stabbing pains. In a few patients the headache has had an almost explosive onset. The pain is usually unilateral, sometimes bilateral, and often localized to the site of the affected arteries in the scalp. The pain persists to some degree throughout the day and is particularly severe at night. It lasts for many months if untreated. The superficial temporal and other scalp arteries are frequently thickened and tender and without pulsation. Jaw claudication and ischemic nodules on the scalp, with ulceration of the overlying skin, have been described in severe cases.
Many of the patients feel generally unwell and have lost weight; some have a low-grade fever and anemia. Usually the sedimentation rate is greatly elevated (>50 mm/h and typically >75 mm/h) but elevation of the C-reactive protein (CRP) level is a more sensitive indicator of this inflammatory condition and is particularly helpful when the sedimentation rate is only mildly elevated. A few patients have a peripheral neutrophilic leukocytosis. As many as 50 percent of patients have generalized aching of proximal limb muscles, reflecting the presence of polymyalgia rheumatica (see Chap. 55, "Polymyalgia Rheumatica").
The importance of early diagnosis relates to the threat of blindness from thrombosis of the ophthalmic or posterior ciliary arteries. This may be preceded by several episodes of amaurosis fugax (transient monocular blindness). Ophthalmoplegia may also occur but is less frequent, and its cause, whether neural or muscular, is not settled. Masticatory claudication is a specific but not particularly sensitive symptom of cranial arteritis. The large intracranial vessels are occasionally affected, thereby causing stroke. Once vision is lost, it is seldom recoverable. For this reason, the earliest suspicion of cranial arteritis should lead to the administration of corticosteroids and then to biopsy of the appropriate scalp artery. Microscopic examination discloses an intense granulomatous or "giant cell" arteritis. If biopsy on one side fails to clarify the situation and there are sound clinical reasons for suspecting the diagnosis, the other side should be sampled. Arteriography of the external carotid artery branches is probably the most sensitive test but is seldom used, because of its relatively higher risk. Ultrasonographic examination of the temporal arteries may display a dark halo and irregularly thickened vessel walls. This technique has not yet been incorporated into the routine evaluation because its sensitivity has not been established; our own experience suggests that it may miss cases, but it could be useful in choosing the site for biopsy of the temporal artery.
The administration of prednisone, 45 to 60 mg/d in single or divided doses over a period of several weeks, is indicated in all cases, with gradual reduction to 10 to 20 mg/d and maintenance at this dosage for several months or years, if necessary, to prevent relapse. The headache can be expected to improve within a day or two of beginning treatment; failure to do so brings the diagnosis into question. When the sedimentation rate or CRP is elevated, its return to normal, usually over months, is a reliable index of therapeutic response. Whether symptoms or the blood tests are a better guide to reducing the steroid dose is unclear, one should probably be cautious in lowering the medication if the erythrocyte sedimentation rate (ESR) and CRP remain high.
Headaches of Pseudotumor Cerebri (Benign or Idiopathic Intracranial Hypertension (see Chap. 30)
The headache of pseudotumor cerebri assumes a variety of forms. Most typical is a feeling of occipital pressure that is greatly worsened by lying down, but many patients have—in addition, or only—headaches of migraine or tension type. Indeed, some of them respond to medications such as propranolol and ergot compounds. None of the proposed mechanisms for pain in pseudotumor cerebri seems to be adequate as an explanation, particularly the idea that cerebral vessels are displaced or compressed, as neither has been demonstrated. It is worth noting that facial pain may also be a feature of the illness, albeit rare. Chapter 30 has a more complete description of the clinical features and treatment.
After successful treatment for pseudotumor, some patients have persistent headaches that have the flavor of migraine.