The modern era of the study of headaches in children and adolescents was initiated by Bille in 1962 (3). He reported data gathered from a population of 9,000 Scandinavian school children. He stated that by age 7, 1.4% of children had experienced episodic migraine, and 2.5% had frequent nonmigrainous headaches. An additional 35% had experienced infrequent headaches of other varieties. By age 15 years, 5.3% of children and adolescents had experienced migraine, 15.7% had experienced frequent nonmigrainous headaches, and 54% had infrequent nonmigrainous headaches. Further communications from Bille suggested that the frequent nonmigrainous headaches are tension-type headaches (13,14). In a study conducted in a prepaid health plan, 80% of all enrolled children were examined for headache during a six-year period. The incidence of new cases of migraine with aura was 6.6/1,000 in males, and the incidence of migraine without aura was 10/1,000 in males. In females, the incidence was 14/1,000 of migraine with aura and 18/1,000 of migraine without aura. The overall prevalence of headaches increases quite strikingly from preschool children through high school. Prevalence of headaches by age 7 ranges from 37% to 51%. From ages 7 to 15, the range increases from 57% to 82%. Recent studies have suggested that the prevalence of headaches in general and migraine in particular may be increasing compared with studies done 10 to 20 years ago. The epidemiology of headaches in children has been reviewed in detail by Lipton (15,16).

Various methods of classifying headaches both in adults and children have been published. Prior to 1988, the classification of headache was not uniform, and diagnostic criteria were not based on operational rules. In 1988, the International Headache Society (IHS) instituted a classification system for headache that has become the standard for headache diagnosis and clinical research (17). The second edition of this classification was published in 2004 (18). Numerous pediatric authors have reviewed the initial IHS classification and have suggested that it is not appropriate for use in children and adolescents. They have
proposed several modifications (19,20). No papers have been published evaluating the second edition of the IHS criteria, as it is too recent. Unfortunately, that edition contains little regarding modifications of the criteria as they apply to children and adolescents.

Prior to and along with these IHS classifications, other clinical methods of classifying pediatric and adolescent headache have been proposed, the most prominent of these by Vahlquist (21) and Prensky and Sommer (22). Winner and Rothner proposed a clinical classification utilizing both the temporal pattern of a child’s headache plotted against its severity over time (7). Five patterns were identified, including acute, acute recurrent, chronic progressive, chronic nonprogressive, and “mixed” or comorbid headaches. A proposed revision of the classification of pediatric headaches is shown in Table 15.1 (23).

An acute headache is a single event with no history of a previous similar event. It may be generalized or localized. It may be associated with neurologic symptoms and signs or seen in the absence of neurologic symptoms or signs. If an acute headache is noted in a critically ill child, a diagnosis needs to be made quickly, and intervention may be lifesaving. The differential diagnosis of acute headaches involves a wide variety of general medical as well as central nervous system etiologies (Table 15.2).

Acute recurrent headaches are periodic headaches that are separated by pain-free intervals. When an acute recurrent headache is associated with nausea, vomiting, photophobia, and phonophobia, the headaches are usually migrainous in nature.

Chronic progressive headaches are headaches that worsen in frequency and severity over time. The progression may occur rapidly or slowly. These headaches may be accompanied by symptoms and signs of increased intracranial pressure or progressive neurologic disease. The neurologic examination is frequently abnormal. An organic process is usually present, and further testing is usually indicated.

Chronic nonprogressive headaches have been referred to under a variety of names, including tension-type headaches, muscle contraction headaches, chronic daily headaches, and chronic nonprogressive headaches. Silberstein and Lipton have recently subdivided chronic daily headaches into chronic tension-type headaches,
hemicrania continua, transformed migraine, and new daily persistent headaches (Table 15.3) (24). Chronic tension-type headaches may occur several times a week, more than 15 days per month, or may be constant. They are usually not associated with symptoms of increased intracranial pressure or progressive neurologic disease. The neurologic examination is normal. Factors relating to school, stress, family dysfunction, and medication overuse are frequently noted. Excessive school absences are common.

Much dispute surrounds the entity of “the mixed headache syndrome” also known as comorbid headaches. Some feel that these are the transformed migraine or chronic migraine as described by other authors. In the latter two entities, the patients begin having episodic migraine, and a number of years later, depending on medication overuse as well as headache frequency, the headaches evolve into a daily vascular headache (24a). Chronic daily headaches are one of the most frequent types of headaches seen in the adolescent population as well as in adults seen at tertiary headaches centers. In our opinion, the condition consists of a combination of acute recurrent headaches, which are migrainous and superimposed on a pattern of chronic nonprogressive daily headaches. At times, it is difficult to differentiate between chronic daily headaches, comorbid headaches, transformed migraine, and chronic migraine. In all of these, however, symptoms of increased intracranial pressure and progressive neurologic disease are absent. The neurologic examination is normal. Laboratory testing is not diagnostic.

Over the years, clinicians have found that the added dimensions of severity and duration as well as the periodicity of the headache are useful in evaluating patients seen in a pediatric or pediatric neurology setting.

If indeed there are various distinct classes of headache, the pathophysiology of each type of headache must be discussed separately. Both extracranial and intracranial structures are sensitive to pain. The sensitive extracranial structures include the skin, subcutaneous tissues, muscles, mucous membranes, teeth, and some of the larger vessels. Pain-sensitive intracranial tissues include the vascular sinuses, larger veins, and dura surrounding these structures and arteries at the base of the brain. Pain from extracranial and intracranial structures in and about the face and from the front half of the skull is mediated via the fifth cranial nerve. Smaller areas are innervated by branches of seventh, ninth, and tenth cranial nerves. Pain from the posterior aspect of skull and upper neck are mediated via the upper cervical nerves. The brain itself and most of the dura, ependyma, and choroid plexus are insensitive to pain. Any process causing inflammation, displacement, irritation, traction, dilation, or physical invasion of any of these pain-sensitive structures will cause pain referred to either the face, top of the head, back of the head, or neck. The pain, at times, may have localizing value. The perception of this pain is certainly modified by the patient’s age, previous experience with pain, and psychologic state, among other factors. The severity of the pain may not be an indication of the severity of the disease.

There have been numerous advances in the understanding of the pathophysiology of migraine. The initial concept was that migraine was secondary to dysfunction of the central nervous system. In 1938, Graham and Wolff proposed the vascular theory of migraine (25). This theory of migraine pathophysiology held that an attack has two phases. The prodromal phase, marked by an aura, was believed to be characterized by vasospasm that induces cerebral ischemia and the various transient focal symptoms that initiate the attack. The second phase, characterized by extracranial vasodilatation, was thought to be responsible for the pulsating headache, which is generally felt in the distribution of the trigeminal nerve and the upper cervical roots (26). It now has been shown that the aura is rarely accompanied by ischemia and that the onset of headache occurs at a time when cortical blood flow is reduced and therefore not caused by vasodilatation (27). The vascular theory has been supplemented by a theory that combines the vascular theory and the neuronal theory and is generally referred to as the trigemino-vascular theory. The elements required to understand this concept include the anatomy of head pain, the physiology and pharmacology of the peripheral branches of the trigeminal nerve and the trigeminal cervical complex, the modification of these systems by brain stem and diencephalic structures, the further connections of these structures to thalamic and cortical areas, and the secondary vascular responses producing plasma protein extravasation.

The trigemino-vascular theory proposes that classical migraine (i.e., migraine preceded by an aura or other focal symptoms) is related to a paroxysmal depolarization of cortical neurons. During the initial phase of the attack, a cortical spreading depression is elicited at the occipital pole of the brain. The term cortical spreading depression is used to describe a depression of spontaneous EEG
and other cortical electrical activities spreading across the cerebral cortical surface in the wake of a variety of noxious stimuli (28). The cortical spreading depression moves anteriorly in the course of the attack at a rate of approximately 2 mm per minute. At the wave front, transient ionic changes cause neurons and glia to depolarize with ensuing neuronal silence. Associated with these changes are dramatic alterations in the ion distribution between intracellular and extracellular departments. These changes are believed to trigger the migraine aura and to induce a 20% to 35% reduction in posterior cerebral cortical blood flow (29). The factors that induce the onset of cortical spreading depression are multiple and also are not well clarified. No doubt, they are both exogenous and endogenous. In part, they include any disturbance of K⁺ homeostasis, genetic predisposition, stress, and dietary factors as well as the antidromic release of vasoactive peptides from the trigeminovascular system (30,31).

Cerebral ischemia is probably the result of arteriolar vasoconstriction and, in most instances, is not of primary importance in the induction of the migraine aura or of focal neurologic symptoms. Generally, oligemia in the posterior part of the brain is the most characteristic alteration of regional cerebral blood flow in attacks of classical migraine (32). Cerebral blood flow in the areas of the brain not affected by cortical spreading depression remains normal. Regional blood flow in the brainstem is increased, however, with maximal increase in the region corresponding to the dorsal raphe nucleus and the locus caeruleus (33). The course followed by the spreading oligemia is independent of vascular patterns and appears to be related to neuronal cytoarchitecture (30). As a rule, the cortical spreading depression stops at the central sulcus. Ventral propagation of the cortical spreading depression to the pain-sensitive meningeal trigeminal fibers that innervate the intracranial and dural blood vessels is believed to induce the headache. Stimulation of the trigeminal sensory neurons results in the release of a number of vasoactive substances. These include vasoactive intestinal peptide, substance P, and calcitonin gene-related peptides. These substances interact with blood vessel walls and induce vascular dilatation, plasma protein extravasation, and platelet activation and induce neurogenic inflammation (34). Neurogenic inflammation sensitizes nerve fibers to the point that they respond to previously innocuous stimuli such as blood vessel pulsations (35). Trigeminal sensitization is probably responsible for the cutaneous allodynia that patients frequently develop in the course of an attack of migraine and whose presence inhibits the effectiveness of triptans (36).

In migraine without an antecedent aura (common migraine), there are no consistent changes in cerebral blood flow (37), and the mechanisms other than cortical spreading depression that are responsible for this form of migraine are poorly understood. They could involve extracranial and intracranial arterial dilatation (38).

In addition to the vascular changes seen in classical migraine, there are a variety of abnormalities in the metabolism and concentrations of neurotransmitters, notably serotonin and its metabolites. At the onset of an attack, serotonin is released from platelets, and during an attack, there is a transient reduction in serotonin turnover (38,39). Between attacks, there appears to be an enhancement in serotonin turnover (40). This observation corresponds to the finding obtained by PET of an increased serotonin synthesis capacity in all brain regions in patients having migraine without aura (40). Of the various serotonin receptors, 5-HT1, 5-HT2, and 5-HT₃ are involved in the pathophysiology of migraine. The 5-HT1 receptors are inhibitory. The postsynaptic 5-HT1_B receptor is located on intracranial blood vessels, whereas the presynaptic receptor, 5-HT1_D, is located on the trigeminal nerve ending. Most of the drugs used for the acute treatment of migraine are 5-HT1_B/5-HT1_D agonists, whereas medications such as propanolol and methysergide are antagonists to the excitatory 5-HT2 receptors (41).

A number of substances can initiate an attack of migraine. These include prostaglandin E₁, tyramine, and phenylethylamine. Tyramine and phenylethylamine can be found in a variety of foods, notably cheese and chocolate, and are responsible for consistently initiating migraine attacks in a significant proportion of adults. In children, however, these amines are of lesser import (42). Considerable evidence suggests that the brain is not the only organ affected during an attack of migraine; alterations in renal functions, particularly polyuria and increased histidine excretion, have been demonstrated (34).

The term tension-type headache implies that the basis of the pain seen in this disorder is related to muscle. This is not really the case. Chronic tension-type headaches have many of the same features of migraine but lack the severe problem with nausea, sensitivity to light and sound, as well as movement. They also lack the usual triggering association such as menses, missing meals, or altering sleep patterns. Possible mechanisms include genetic aspects, muscle mechanisms, and central and/or peripheral sensitization (43,44). A few studies have suggested that there are genetic factors that play a role in chronic tension-type headaches (45). Stress also is identified as a trigger, but neither of these two triggers helps us to understand the basic pathophysiology better. Various experiments of injecting substances into muscle have reproduced pain, as has direct electrical stimulation of muscles. However, these types of pain do not seem to produce the same dull bifrontal headache that is seen in chronic tension-type headaches. Muscles that seem to be more sensitive to touch than normal controls have not been found to be different when studied neurophysiologically. Whether “anoxia” or
nitric oxide generation plays a role in causing muscle pain remains to be proven. EMG data is controversial. Initial perceptions were that muscle contraction generates the pain seen in tension-type headaches. Although EMG activity tends to be higher in patients than in controls, there is no increased EMG activity during a headache, compared with EMG activity in the absence of a headache (46,47). Until the clinical definition is better clarified, the pathophysiology of chronic tension-type headaches will not be elucidated.

Cluster headaches are a clinically well-defined disorder. The clinical condition closely resembles other disorders, such as paroxysmal hemicrania or SUNCT (short-lasting unilateral neuralgiform headache with conjunctival injection and tearing). There are three major aspects to the pathophysiology of this disorder—the trigeminal distribution of the pain, the autonomic features associated with the pain, and the episodic pattern of the attacks (48). The majority of patients affected are males, and the neuroendocrine abnormalities of testosterone levels altered during cluster headaches have been known for over 30 years (49). Other interesting observations relating to cortisol, growth hormone, and prolactin also have been published (49). The hypothalamus as well has been implicated, since these headaches characteristically awaken patients at the same time each night. Imaging studies indicate increased activation of the anterior cingula and the contralateral frontal cortex, insula, and thalamus (50). The ipsilateral hypothalamic gray matter becomes activated during cluster headaches. Vascular changes seem to be secondary to the above. Additionally, there is some initial data suggesting that in a small percentage of cluster patients, genetics are implicated (48).

The familial transmission of migraine has been known for centuries. Although in the past the condition was considered to be transmitted as an autosomal dominant disorder, current genetic data do not support this pattern. Almost all genetic studies have been confounded by the high prevalence of migraine in the general population, which facilitates a chance familial occurrence. In addition, inclusion of a family history of migraine as a diagnostic criterion, differences in migraine case definition, and variation in case ascertainment (referral to a clinic vs. a population survey) all have hindered valid genetic studies. However, over the last few years, several studies have shed light on the genetics of the disorder. Twin studies have supported a strong genetic component in the etiology of migraine, with a significantly higher concordance rate among monozygotic twins as compared with dizygotic twins (51,52,53). Subjects with classical migraine are more likely to have first-degree relatives with classical migraine than with common migraine, and vice versa. From data such as these, it has become evident that migraine with aura (classical migraine) and migraine without aura (common migraine) are genetically distinct entities (54). Although some authors have postulated an autosomal recessive model with reduced penetrance, it is more likely that both migraine entities have a multifactorial inheritance and that there is no evidence for any specific Mendelian transmission (55,56,57). In any case, the frequency of migraine indicates that in the future, multiple genes will almost certainly be implicated. Familial hemiplegic migraine has offered the most fruitful lines of genetic research. This condition is covered in another section of this chapter.

An excellent review of the clinical presentations can be found in the monograph by Winner and Rothner (7).

The migraine syndrome is the classic example of an acute recurrent headache. Episodic migraine is characterized by episodic, periodic, paroxysmal attacks of throbbing headache, which may be unilateral or bilateral. The attacks are separated by pain-free intervals. They are often preceded by pallor and behavioral change and are often associated with decreased appetite, nausea, vomiting, phonophobia, and photophobia. They are frequently relieved by sleep.

Migraine is relatively common in children. A survey published in 1997 of 3- to 11-year-old children in a British general practice using a questionnaire and a structured interview disclosed that depending on the diagnostic criteria, 3.7% to 4.9% of children experienced migraine. Of these, 1.5% had migraine with aura (71).

Migraine begins surprisingly early in life. Initial complaints are paroxysmal and recurrent abdominal pain, restlessness, head banging, or sudden alterations in personality. A history of motion sickness or carsickness can be elicited in approximately two-thirds of patients (72). Because these symptoms are nonspecific, the diagnosis of migraine is generally not made until the child is old enough to relate the symptoms. Approximately one out of five patients has the first attack before age 5 years (73). Prior to puberty, boys are affected almost twice as often as girls (74).

As already noted, headaches are characteristically paroxysmal and are separated by symptom-free intervals. Commonly, particularly in teenagers, an attack begins in the early morning hours, often awakening the child. In younger children, the attacks tend to start in the afternoon. In the classic form of migraine, many attacks of headache are preceded by an aura. The most common symptoms in children involve nausea, vomiting, abdominal pain, and disturbances of vision. Some of the older children describe scintillating scotomata moving across one or both visual fields, and in adults, visual symptoms are the most common manifestation of the aura (75). Vision is blurred, and the child can have a transient hemianopsia or even complete blindness in one eye (amaurosis fugax). Both can terminate with the onset of contralateral headache or can last for several days unaccompanied by head pain (76). A family history of the disorder can be elicited in over one-half of the patients and was found in 72% in the data compiled by Prensky (74).

Other symptoms preceding the headache include numbness and tingling in one arm or over the entire side, hemiplegia, aphasia, or apraxia (77). In most instances, symptoms appearing during the preheadache phase are completely reversible, but in some children, function returns slowly, and an occasional case of persistent hemianopsia or hemiparesis has been reported (78,79). Some of the latter cases undoubtedly were subjects with familial hemiplegic migraine.

In younger children, migraine headache is bifrontal or poorly localized and is almost invariably accompanied by pallor, nausea, and vomiting (5). Occasionally, vomiting can be sufficiently intense and prolonged to induce acidosis and mimic cyclic vomiting. The relationship between migraine and cyclic vomiting is not clear. Some patients with cyclic vomiting have a strong family history of migraine, and a significant proportion develop migraine in adult life (80,81). An attack of migraine lasts anywhere from half an hour to several days.

In a small proportion of children, focal neurologic symptoms that are present during an attack persist beyond the headache phase. The term complicated migraine has been applied to forms of migraine that include those of the
hemiplegic and basilar artery migraine. Ophthalmoplegic migraine is no longer considered a complicated form of migraine. The authors feel that any attack of migraine associated with even transient neurologic disturbances requires further thought and investigation. The neurologic syndromes associated with migraine are defined by their vascular territories. In the majority of patients, complete recovery is the rule. Certainly, any patient left with neurologic deficits requires further investigation if the initial investigation has not been carried out. Structural abnormalities may on occasion mimic migraine. It is best to think of a patient with migraine and neurologic features as having an underlying neurologic disorder until proven otherwise. A noninvasive evaluation that includes MRI should be utilized prior to the diagnosis of complicated migraine. Invasive angiographic studies are only uncommonly needed.