Walsh and Hoyt's Clinical Neuro-Ophthalmology

Headache and Facial Pain

"Headache and facial pain are common complaints and represent a diverse range of etiologies, from benign to lifeand vision-threatening."

"Headache and facial pain are common complaints and represent a diverse range of etiologies, from benign to lifeand vision-threatening. Headache, migraine in particular, has been described in the popular and medical literature for over 3000 years (1). Trepanation, a sign of neurosurgery, has been seen on Neolithic skulls dating from 7000 B.C. (Fig. 26.1). The Ebers papyrus, an ancient Egyptian prescription dating back to 1200 B.C., mentions migraine, neuralgia, and shooting head pains, and is thought to be based on earlier medical documents from 1550 B.C. (2). Indeed, it is estimated that over 90% of individuals have noted at least one headache over their entire life (3). Patients with headache may present initially to a neurologist or ophthalmologist, and are often misdiagnosed initially. Since correct diagnosis facilitates appropriate treatment, it is important for the clinician to be familiar with common causes of head and facial pain. This chapter will focus primarily on those causes of headache and facial pain most relevant to neuro-ophthalmologists."

CHAPTER 26 Headache and Facial Pain Gregory P. Van Stavern APPROACH TO HEADACHE AND FACIAL PAIN Trigeminal Autonomic Cephalgias CLASSIFICATION OF HEADACHE: INTERNATIONAL Other Short-Lasting Primary Headaches HEADACHE SOCIETY CRITERIA SECONDARY HEADACHE SYNDROMES IMPACT OF HEADACHE Head Trauma and Headache NEUROANATOMY AND NEUROPHYSIOLOGY OF PAIN Headache Attributable to Cranial and Cervical Vascular Disorders Pain Terminology Headache Attributable to Intracranial Infections Initiation Phase Headache Attributable to Nonvascular Intracranial Disorders Maintenance of Pain Headache Attributable to Disorders of Homeostasis PRIMARY HEADACHE SYNDROMES Headache Attributable to Disorders of the Eyes Migraine Headache Attributable to a Substance or Its Withdrawal Tension-Type Headache Cranial Neuralgias and Central Causes of Facial Pain Headache and facial pain are common complaints and represent a diverse range of etiologies, from benign to life-and vision-threatening. Headache, migraine in particular, has been described in the popular and medical literature for over 3000 years (1). Trepanation, a sign of neurosurgery, has been seen on Neolithic skulls dating from 7000 B.C. (Fig. 26.1). The Ebers papyrus, an ancient Egyptian prescription dating back to 1200 B.C., mentions migraine, neuralgia, and shoot-ing head pains, and is thought to be based on earlier medical APPROACH TO HEADACHE AND FACIAL PAIN Headache and facial pain result from disorders that affect the pain-sensitive structures in the head and neck, such as meninges, blood vessels, and muscles. Pain-sensitive struc-tures responsible for head and facial pain are listed in Table 26.1. A history is the first and most important step in the evalu-ation of head and facial pain; the examination and any ancil-lary tests serve to confirm or exclude what is already sus-pected. In many patients with headache and facial pain, the general neurologic and ophthalmologic examination is nor-mal, and the correct diagnosis rests upon a thorough and accurate history. Most, but by no means all, of the more serious and life-threatening causes of headache present with acute onset of pain: a patient with his or her ‘‘first or worst'' headache has a greater likelihood of harboring an ominous cause for the head pain than someone with a 10-year history 1275 documents from 1550 B.C. (2). Indeed, it is estimated that over 90% of individuals have noted at least one headache over their entire life (3). Patients with headache may present initially to a neurologist or ophthalmologist, and are often misdiagnosed initially. Since correct diagnosis facilitates ap-propriate treatment, it is important for the clinician to be familiar with common causes of head and facial pain. This chapter will focus primarily on those causes of headache and facial pain most relevant to neuro-ophthalmologists. of recurrent, low-grade headache that has not changed in character. Although a chronic and recurrent headache is more likely to represent a benign condition, patients with a headache distinct from previous headaches (in terms of loca-tion or quality of pain) are also more likely to harbor a seri-ous underlying illness. The location of the pain may prove helpful. Unilateral headache is an invariable feature of cluster headache and occurs with most migraine attacks. Patients with tension-type headaches generally describe bilateral pain. In patients with ocular pain, a primary ophthalmic etiology should be excluded, while paranasal pain localized to the sinuses might suggest sinus obstruction. A new temporal headache in an elderly patient might raise suspicion of giant cell arteritis (GCA). Figure 26.2 depicts regional pain in common pri-mary headache syndromes. However, since most head and 1276 CLINICAL NEURO-OPHTHALMOLOGY Figure26.1. Neolithic skull showing trepanation hole (ca. 7000 B.C.) (Courtesy of Nationalmuseet, Copenhagen. From Headache in Clinical Practice. Oxford, UK: Isis Medical Media, 1998.) facial pain is ultimately mediated by the trigeminal nerve, referred pain may result in false localization. A complete headache history should include a description of the quality of the pain. Migraine headache is typically reported as throbbing; tension-type headache is often de- Figure26.2. Sites of common head and facial pain syndromes. (From Lance JW. Migraine and Other Headaches. New York: Charles Scribner's Sons.) Table 26.1 Pain-Sensitive Structures of the Head and Neck Extracranial Skin and blood vessels of the scalp Head and neck muscles Second and third cervical nerves Periosteum of the skull Eyes, ears, teeth, sinuses, oropharynx Mucous membranes of nasal cavity Intracranial Dura mater Anterior and middle meningeal arteries Trigeminal (V), glossopharyngeal (IX), and vagus (X) nerves Proximal portions of internal carotid artery and branches near the circle of Willis Periaqueductal gray matter Sensory nuclei of the thalamus scribed as dull and aching, with a band-like sensation around the head. The pain produced by intracranial mass lesions is often dull and steady. A sharp, lancinating quality suggests neuralgic pain. The tempo and evolution of the pain may suggest a specific diagnosis, since some headache syn-dromes have fairly characteristic temporal patterns (Fig. 26.3).The onset and duration of the attacks, including time HEADACHE AND FACIAL PAIN 1277 Figure26.3. Temporal patterns of headache. (From Lance JW. Mechanism and Management of Headache. Ed 5. Oxford, UK: Butterworth-Heinemann, 1993.) to peak intensity, should be determined. The frequency of attacks should be ascertained; a change in frequency of a preexisting headache raises concern about a secondary head-ache syndrome superimposed upon previous primary head-aches. Exacerbating and relieving factors may provide clues about etiology. Migraine is often provoked by certain foods, worsened by routine physical activity, and may be worse during menses. A positional headache might suggest either increased or decreased intracranial pressure. Pain provoked by chewing or eating might raise suspicion of giant cell arter-itis or temporomandibular joint dysfunction. Precipitation of headache with alcohol is common with cluster headache. Migraine headaches are frequently relieved by darkness and sleep. Associated focal neurologic symptoms which do not meet criteria for migraine with aura might merit further in-vestigation. The general medical history is valuable, as well. Previous head injury might suggest post-traumatic headache. Medical co-morbidities such as hypertension may be rele-vant when choosing pharmacologic treatments. Family his-tory should be reviewed, as some headache syndromes have a genetic component. A detailed medication history is impor-tant, since analgesic overuse may complicate many chronic headaches. Medication history also may help guide future medical treatment. The general examination may help identify an underly-ing systemic illness. Fever might suggest an intracranial or meningeal infection, but headache can accompany nearly any systemic infectious illness. Inspection of the skin may reveal cutaneous lesions associated with a particular diagno-sis (e.g., vesicular lesions in patients with herpes zoster). Blood pressure measurement is important. Although chronic hypertension does not cause headache, an acute rise in arte-rial pressure might cause sudden head pain. Stroke and sub-arachnoid hemorrhage (both of which cause headaches) are both associated with an acute rise in blood pressure. Palpa-tion of the temporal arteries should be performed in all cases where giant cell arteritis is suspected: focal tenderness over the artery, a palpable cord, or obliteration of the pulse in-crease the likelihood of the disease. Meningeal signs should be sought in patients with an acute onset of headache. Recall that meningeal irritation causes nuchal rigidity in the ante-rior- posterior rotation, while cervical spine disorders restrict movement in all directions. The neurologic and neuro-ophthalmic examination is aimed primarily at confirming or excluding a focal deficit, or findings localizable to a specific cranial structure. Such finding would then prompt appropriate further evaluation. The neuro-ophthalmologic examination is essential in the evaluation of a patient with headache or facial pain. Fun-duscopic examination should be performed systematically looking for papilledema; the finding of a Horner syndrome in a patient with unilateral facial or head pain may suggest cluster headache or internal carotid dissection; a 6th nerve palsy or comitant esotropia may raise the possibility of raised intracranial pressure. Finally, ocular causes of facial pain should not be overlooked and a red eye or orbital symptoms should prompt a detailed ophthalmologic exami-nation. In many patients, the examination is entirely normal, and the physician must decide whether or not to obtain neuroim-aging. Since computerized tomography (CT) and magnetic resonance imaging (MRI) are noninvasive and have become widely accessible, many patients with headache will be im-aged, whether or not there are clear indications. This is, in part, fostered by the medicolegal climate in the United States. However, neuroimaging is expensive and inconven-ient for many patients. In addition, if neuroimaging is re-quired, it is preferable to order the correct study the first time (e.g., MRI or MRA rather than CT). The American Academy of Neurology position paper, based on an analysis of 16 CT and MRI studies, concluded that the routine use of CT or MRI was not warranted in patients whose head- 1278 CLINICAL NEURO-OPHTHALMOLOGY aches fit a broad definition of recurrent migraine and who have had no recent change in headache pattern and have no focal neurologic deficits on examination (4). The U.S. Headache Consortium guidelines evaluated the utility of neuroimaging studies in patients with headache. Significant intracranial lesions were found at a rate of only 0.18% in patients with migraine and normal neurologic examinations (5). Some patients with migraine may have non-specific ab-normalities on MRI (6,7). Igarashi and colleagues (6) found that 31% of migraineurs had small foci of T2 hyperintensity on MRI; the signal changes were most commonly seen in the centrum semiovale and frontal white matter (Fig. 26.4). A significantly lower number of matched controls without migraine had similar signal changes on brain MRI. There was no correlation between MRI changes and migraine type; ergotamine use; or frequency, duration, or intensity of mi-graine. These MRI findings are of uncertain clinical signifi-cance. Forsyth and Posner (8) evaluated 111 patients with pri-mary and metastatic brain tumors, and identified headache in 48%. Many of these patients had headaches that were similar to previous headaches, but now were associated with seizures, prolonged nausea, confusion, or other neurologic findings. The authors concluded that a change in headache symptoms is an indication for neuroimaging. In most cases, MRI is superior to CT scan for evaluation. There are a num-ber of disorders causing headache that may be missed by CT scanning, particularly if contrast is not administered (Table 26.2). There are few published studies evaluating the value of neuroimaging in some of the less common headache disor-ders. The decision to image in those circumstances should Figure26.4. MRI of a 53-year-old woman with a long history of migraine. FLAIR sequences. The size and location of the white matter lesions are relatively nonspecific and are atypical for demyelinating disease. Table 26.2 Lesions that May Be Missed by CT Scan Low-grade glial tumors Vascular malformations (particularly in noncontrast CT) Venous sinus thrombosis Cavernous sinus and sellar lesions Meningeal disease be made on a case by case basis. Patients with chronic eye pain and a normal neuro-ophthalmic examination are a too-common source of referral to ophthalmologists and neuro-ophthalmologists. These patients are often classified as hav-ing ‘‘atypical facial pain'' and in many cases have a substan-tial component of analgesic overuse and rebound. The value of imaging in these patients is uncertain. Golnik et al. (9) reviewed 101 patients with unilateral eye and facial pain and noted no findings on examination that would account for the pain. Neuroimaging was normal in 93% of patients; in those with abnormal imaging it was unclear whether the pain was attributable to the lesions found. They concluded that the yield of neuroimaging in a patient with unilateral eye pain and a normal examination is low. The decision to image should be made on a case by case basis. Table 26.3 lists reasonable indications for neuroimaging. For many patients the etiology of headache and facial pain can usually be ascertained after a careful history and examination. An awareness of relevant neuroanatomy as well as typical presentation for common headache syn-dromes is essential. If neuroimaging is indicated, the type if image ordered should be the one most likely to confirm or exclude certain diagnoses. For example, CT scanning has high sensitivity for bony lesions and acute hemorrhage and is the test of choice for acute head trauma and neurologic deficits with abrupt onset. MRI has much higher sensitivity Table 26.3 Indications for Neuroimaging in Patients with Head and Facial Pain History The first or worst headache of the patient's life, particularly if abrupt onset Change in frequency, severity, or clinical features of headache attack Neurological symptoms that do not meet criteria for migraine with aura Hemicrania that is always on the same side and is associated with contralateral neurologic symptoms Positional headache Lack of improvement with conventional therapy Examination Localizable deficits on neurologic or neuro-ophthalmologic exam Persistent neurologic or neuro-ophthalmologic deficits HEADACHE AND FACIAL PAIN 1279 Figure26.5. Algorithm for management of patients with headache. for posterior fossa, cavernous sinus, and sella turcica le- arterial (e.g., dissection, aneurysm) or venous (e.g., sions. The MRI may miss orbital pathology unless fat- cerebral venous thrombosis) disease is suspected. Figure suppressed orbital sequences (in which the bright T1 signal 26.5 presents a reasonable algorithm for evaluating patients from orbital fat is eliminated) are specifically ordered. presenting with head and facial pain. It is not intended Magnetic resonance angiography and venography are to be all-inclusive, but rather provide a framework for specialized MRI sequences that may be ordered when managing patients with headache. 1280 CLINICAL NEURO-OPHTHALMOLOGY CLASSIFICATION OF HEADACHE: INTERNATIONAL HEADACHE SOCIETY CRITERIA Headache may occur in isolation or as part of a symptom complex-the primary headache syndromes, or may be part of an underlying condition, such as brain tumor or stroke -the secondary headache syndromes. Management of sec-ondary headache syndromes is largely determined by the underlying structural or metabolic disorder responsible for the generation of pain. Management of primary headache syndromes, however, is dependent upon accurate diagnosis and classification, which then serves to guide therapy. Prior to 1988, there were no standardized headache classification systems with operational rules and uniform nomenclature. In 1988, the International Headache Society (IHS) instituted a classification system that has become the standard for headache diagnosis (10). The IHScriteria have received broad international support, and the principles of the system have been included in the International Classification of Dis-eases (ICD). The IHScriteria have established uniform ter-minology and consistent diagnostic criteria for a range of headache disorders; in turn, this has facilitated epidemiologi-cal studies and clinical trials that provide the basis for current research and treatment guidelines. The second edition of the IMPACT OF HEADACHE Headache is a public health problem of tremendous scope and impact. A U.S. survey of 40,000 households (112,000 persons) found that 5.5 days of restricted activity per 100 person-years were due to headache (12). In another study, 8% of males and 14% of females missed all or part of a day at work in the 4-week period prior to the interview (13). Headache also results in frequent utilization of emergency rooms and urgent care centers. Despite this, many headache syndromes (including migraine) remain underdiagnosed. NEUROANATOMY AND NEUROPHYSIOLOGY OF PAIN The International Society for the Study of Pain defines pain as ‘‘an unpleasant sensory and emotional experience associated with actual or potential tissue damage'' (15). Pain is therefore multidimensional, and may be good or bad, nec-essary or unnecessary, and is dependent upon self-report of the experience. The relationship between disease (inflamma-tion, tissue damage, nerve damage) and central processing determines a wide range of pain symptoms. Peripheral and central pain pathways have plasticity, and can be modified both functionally and structurally (16). These concepts pro-vide insight both into the underlying mechanisms of pain syndromes, as well as the rationale for therapeutic interven-tions. PAIN TERMINOLOGY Before discussing the pathways involved in processing pain, and the variety of pain syndromes, it is helpful to re-view some of the terminology used in pain literature. Re-ferred pain is pain in an area far removed from the site of tissue injury. Phantom pain refers to pain in a part of the IHSclassification was recently completed and follows the same principles as the first edition (11). Several headache syndromes that did not have sufficient valid evidence (e.g., carotidynia) have been eliminated, and new headache syn-dromes (e.g., hemicrania continua) have been added. As in the first edition, the classification draws upon all forms of available evidence, including clinical description, longitudi-nal and epidemiological studies, treatment results, genetics, neuroimaging, and pathophysiology. That the classifications remain largely descriptive, in part reflects the subjective na-ture of pain symptoms. As research identifies more specific, objective markers for pain disorders, the IHScriteria will likely incorporate these findings in future editions. At this point, however, the IHSclassification system remains an invaluable source of information and the primary reference for any physician evaluating and managing patients with head and facial pain. Table 26.4 lists the headaches that will be covered in this chapter; they are categorized according to the IHSclassification system. The headaches most rele-vant to neuro-ophthalmologists are highlighted and will be covered in greater detail than the other headaches listed. Data from the American Migraine Study II show that fewer than half (48%) of patients meeting IHScriteria for migraine reported a physician diagnosis of migraine (14). It is therefore important for all physicians encountering patients with headache to be familiar with diagnostic criteria, ex-clude secondary headache syndromes (through history, exam, diagnostic studies), and either begin management or refer the patient to a physician who can begin appropriate treatment. body that has been surgically removed or is congenitally absent. Allodynia is a nonnoxious stimulus that is perceived as painful (e.g., light touch producing pain). Sensitization occurs when a receptor responds to a stimulus in a more intense fashion than expected, or to a stimulus to which it would not normally respond. Hyperalgesia is exacerbated pain produced by a noxious stimulus (e.g., mild pinprick causing severe pain out of proportion to the stimulus). Hy-perpathia refers to intense pain with repetitive stimuli. An-esthesia refers to complete loss of sensation. Paresthesias are abnormal sensations, such as burning, tingling, or formi-cation, that are not typically painful. A nociceptor is a ner-vous system receptor capable of distinguishing between a noxious and innocuous stimulus. Head and facial pain may be broadly classified as primary or secondary. Primary headaches may be severe and inca-pacitating but do not cause death or permanent neurologic deficit, and no identifiable etiology is identified. Secondary headaches may be caused by a wide range of extracranial and intracranial pathology, and correction of the underlying HEADACHE AND FACIAL PAIN 1281 Table 26.4 Causes of Head and Facial Pain PRIMARY HEADACHES 1. Migraine a. Migraine without aura b. Migraine with aura i. Typical aura with migraine headache ii. Typical aura without headache iii. Familial hemiplegic migraine iv. Basilar type migraine c. Retinal migraine d. Persistent aura without infarction 2. Tension-type headache a. Frequent episodic tension-type headache b. Chronic tension-type headache 3. Cluster headache and other trigeminal-autonomic cephalgias a. Cluster headache b. Paroxysmal hemicrania i. Episodic paroxysmal hemicrania ii. Chronic paroxysmal hemicrania c. Short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT) 4. Primary stabbing headache 5. Hemicrania continua SECONDARY HEADACHES 1. Headaches attributed to head and/or neck trauma a. Acute post-traumatic headache b. Chronic post-traumatic headache 2. Headache attributed to cranial and cervical vascular disorders a. Ischemic stroke and transient ischemic attacks b. Subarachnoid hemorrhage c. Subdural hematoma d. Epidural hematoma e. Unruptured vascular malformation f. Arterial dissection g. Pituitary apoplexy h. Giant cell arteritis 3. Headache attributed to nonvascular intracranial disorders a. High cerebrospinal fluid pressure i. Idiopathic intracranial hypertension b. Low cerebrospinal fluid pressure c. Intracranial neoplasm 4. Headache attributed to infection 5. Headache attributed to disorders of homeostasis a. Systemic hypertension b. High-altitude headache 6. Headache of facial pain attributed to disorders of cranium, neck, eyes, ears, nose, sinuses, teeth, mouth, or other facial or cranial structures a. Acute glaucoma b. Refractive errors c. Ocular inflammatory disorders 7. Headache attributed to a substance or its withdrawal a. Medication overuse headache 8. Cranial neuralgias and central causes of facial pain a. Trigeminal neuralgia b. Occipital neuralgia c. Supraorbital neuralgia d. Herpes zoster e. Ophthalmoplegic migraine f. Idiopathic facial pain Disorders most relevant to neuro-ophthalmologists and ophthalmologists are in boldface. abnormality typically results in relief of pain. All forms of head and facial pain reflect activation of small-caliber noci-ceptive neurons projecting from the trigeminal and upper cervical dorsal root ganglia to intracranial vascular and me-ningeal structures. The mechanism by which these nocicep-tive neurons are initially depolarized is the defining differ-ence between primary and secondary headache syndromes. Once activation has occurred, the underlying principles gov-erning the generation and maintenance of pain is likely to be similar for primary and secondary headaches. The brain is an insensate organ. During neurosurgical pro-cedures, Penfield (17) demonstrated that stimulation of the brain parenchyma in awake patients caused no pain. The pia, dura, and extracranial blood vessels are innervated by an adventitial plexus comprised of the trigeminal and upper cervical dorsal root ganglia fibers. These pathways mediate all head pain. Pain-sensitive structures of the head and face include the skin and blood vessels of the scalp, the dura, the venous sinuses, the arteries, and the sensory fibers of the fifth, ninth, and tenth nerves (Table 26.1). The primary noci-ceptive neurons project unmyelinated C fibers which, when activated, transmit nociceptive signals from perivascular ter-minals through the trigeminal ganglia and synapse centrally on second-order neurons in the trigeminal nucleus caudalis (Table 26.5). The primary afferents are mainly gluta-matergic, but they also store substance P, calcitonin G-re-lated peptide (CGRP), and neurokinin A (NKA), all of which may be involved in the modulation of pain (18,19). Trigemi-novascular neurons terminate within the trigeminal nucleus caudalis (TNC). Activity within this nucleus is modulated by projections from several sites. Second-order neurons carry nociceptive information from the TNC to numerous subcor-tical sites, including the cerebellum, the thalamus, as well as other areas of the brain responsible for the emotional response to pain (20,21). INITIATION PHASE The initiation of trigeminal and upper cervical nociceptive neurons in secondary headache syndromes is relatively well understood, and may occur by a variety of mechanisms: 1. Mechanical traction (e.g., tumors, vascular malforma-tions). 2. Chemical activation (e.g., subarachnoid hemorrhage, pu-rulent infections). 3. Inflammatory activation (e.g. autoimmune vasculitis). 4. Direct injury (e.g., head trauma, intracranial surgery). Table 26.5 Processing of Craniovascular Pain Order Anatomic Structures Substructures/Comments 1st 2nd 3rd Final Trigeminal ganglion Trigemino-cervical complex Thalamus Cortex Located in middle cranial fossa Trigeminal nucleus caudalis and dorsal horns of C1 and C2 Ventrobasal complex Processing uncertain 1282 CLINICAL NEURO-OPHTHALMOLOGY 5. Exposure to or withdrawal from toxins (e.g., nitroglycer-ine, caffeine). 6. Metabolic disturbances (e.g., hypoglycemia). Secondary headache syndromes often have focal neuro-logic findings that are far more compelling than the head-ache. Intracranial neoplasms and infectious meningitis fre-quently present with seizures, and the headache is a minor component of the clinical presentation. However, headache may be the inaugural manifestation of some disorders, such as cerebral venous thrombosis, and focal findings may be delayed for days or weeks (22). The mechanism of initiation and activation of nociceptive neurons and afferent fibers in primary headache syndromes (e.g., migraine, cluster) remains poorly understood. Since the diagnoses are descriptive, these syndromes may repre-sent final common pathways for a diverse group of pro-cesses, some acquired, some genetically determined. MAINTENANCE OF PAIN In secondary headache syndromes, the system remains activated by a persistent stimulus, such as traction on pain-sensitive structures by a tumor. In primary headache syn-dromes such as migraine, the mechanism by which the pain persists and intensifies long after activation has yet to be fully elucidated. There are data suggesting that more than one mechanism may be involved. Peripheral Sensitization Sustained peripheral activity can induce significant func-tional and structural changes in the CNS. Activation of tri-geminal nociceptive neurons mediates the central transmis-sion of pain, but also results in the release of vasoactive neuropeptides, including substance P, CGRP, and NKA. These substances induce a sterile inflammatory response, which involves leakage of plasma and plasma protein from small vessels into surrounding tissue, vasodilation, and acti-vation of mast cells (23,24). This process has been termed neurogenic inflammation, which under normal conditions minimizes real or threatened tissue injury. This response is felt to be associated with a lowered threshold for subsequent reactivation and may be a method by which headache is amplified and maintained. The release of CGRP from tri-geminal sensory afferents causes vasodilation and plasma extravasation from dural vessels. Whether such effects occur in humans during primary headaches (such as migraine) is PRIMARY HEADACHE SYNDROMES MIGRAINE Migraine is derived from the Greek word hemicrania, in-troduced by Galen in approximately 200 B.C. (2). Over the years, a variety of popular names for this disorder have evolved, including sick headache, blind headache, and bil-ious headache. Because of his classic descriptions, Aretaeus of Cappadocia (second century A.D.) is credited as the dis-coverer of migraine headaches. Migraines have appeared in uncertain. However, intravenous infusion of CGRP into sus-ceptible individuals elicited migraine-like headaches (25). Central Sensitization Recent studies suggests that chemical irritation of the me-ninges causes the trigeminovascular fibers innervating the dura and central trigeminal neurons to respond to low-inten-sity stimuli that previously induced minimal or no response (26). This may also reflect structural changes as well. For example, constant firing of A-beta nociceptive fibers may lead to reorganization of A-beta terminals in the dorsal horn of the spinal cord (16). Central sensitization may account for the severe prolonged pain of certain headaches. Burstein and colleagues (26) assessed allodynia during migraine at-tack and found sensitization of the trigeminal system at a second- or third-order neuron level. The phenomenon of wind-up is well described in animal research and refers to the increase in dorsal horn nociceptive neuron responsive-ness (in both magnitude and duration) with each subsequent stimulus above a certain frequency (27). Wind-up is sensi-tive to N-methyl-D-aspartate (NMDA) receptor antagonists, and a potential mechanism in second-order neuron sensitiza-tion may be the removal of magnesium gating at NMDA receptors. This suggests that pharmacologic strategies aimed at NMDA receptors might benefit some primary headache syndromes (2). The interplay between peripheral and central activation, amplification, and sensitization in primary headache syn-dromes remains poorly understood. The fact that headache occurs ipsilateral to a dysfunctional hemisphere (as occurs during migraine aura) suggests that the brain can activate or sensitize, directly or indirectly, meningeal nociceptive neu-rons. Further, limbic prodromal symptoms such as mood change and altered appetite suggest that brain dysfunction may precede and possibly provoke headache in certain set-tings. Recent observations using positron emission tomogra-phy (PET) have implied a primary ‘‘generator'' role for ros-tral brainstem structures in patients with migraine (23). The structures that seemed most involved were the periaqueduc-tal gray, the dorsal raphe nucleus, and locus ceruleus. It is also possible that these centers are responsible for modulat-ing the flow of pain impulses, rather than primary generation of pain. It is likely that brainstem centers play a key role in the generation or modulation of pain, and current models of headache and facial pain provide a framework for further investigation. Understanding the pathophysiology may help guide treatments designed to prevent or relieve pain. popular literature for centuries. Shakespeare describes mi-graine treatment in Othello and King John (28,29). Lewis Carroll described migrainous phenomena in Alice in Won-derland and Through the Looking Glass; these included cen-tral scotoma, tunnel vision, distortions in body image, and visual hallucinations (30). It is therefore perplexing that, even today, migraine remains underdiagnosed and under-treated (14,31). There are a number of contributing factors. Some health care providers may not devote time and re- HEADACHE AND FACIAL PAIN 1283 sources to identifying migraine, particularly given the lim-ited time available to evaluate patients in a nonspecialty clinic. Some patients with migraine may fail to seek medical attention. Recent data suggest that the presence of concomi-tant headache types may interfere with diagnosis as well (14,31). Epidemiology and Impact of Migraine Few studies of migraine incidence have been performed, and none has systematically followed a large sample of head-ache- free individuals across a broad range of ages to ascer-tain new cases of IHS-defined migraine. Indeed, it is chal-lenging to estimate the incidence of a chronic disorder with episodic manifestations. Stewart (32) estimated migraine in-cidence using prevalence data. In males, the peak incidence of migraine with aura was 6.6/1000 person-years at 5 years of age; migraine without aura in men peaked at 10/1000 person-years between the ages of 10 and 11 years. In women, the incidence of migraine with aura peaked between ages 12 and 13 (14.1/1000 person-years); migraine without aura peaked between ages 14 and 17 (18.9/1000 person-years). Migraine prevalence varies by age, gender, race, geog-raphy, and socioeconomic status (2). Prior to puberty, mi-graine prevalence is higher in boys than in girls. As adoles-cence approaches, prevalence increases more rapidly in girls than boys. Prevalence increases until age 40, after which it declines (Fig. 26.6). Migraine prevalence is highest between the ages 25 and 55 (33). Race and geographic region contrib-ute to migraine prevalence. In one population-based study, prevalence of migraine was lowest in Asian-Americans, in-termediate in African-Americans, and highest among Cauca-sians (34). Several studies have suggested that migraine prevalence is inversely related to socioeconomic status (33,35). This contradicts prior popular beliefs that migraine was more prevalent among those with higher socioeconomic status. However, accurate medical diagnosis of migraine is more common among high-income group, which may have given rise to this misperception. Figure26.6. Migraine prevalence by age. (From Silberstein SD, Lipton RB. Epidemiol-ogy of migraine. Neuroepidemiology 1993; 12 179-194.) The economic impact of migraine includes direct costs (the expense of diagnosing and treating a particular disorder) as well as indirect costs (the aggregate economic effects of migraine on productivity at work, at home, and in other roles). It is estimated that migraine headache costs U.S. em-ployers approximately $13 billion per year as a result of missed work and reduced productivity (36). Genetics of Migraine Large-scale epidemiologic studies have confirmed the contribution of genetic factors in migraine. In the past, popu-lation- based studies were limited by inaccurate diagnosis of migraine in cohorts. The advent of the IHScriteria in 1988 allowed for a more precise definition of migraine. Even so, misdiagnosis and underdiagnosis remain a potential risk, and some patients who do not report migraine at the time of the study may develop symptoms later in life. A study of Danish migraineurs found that first-degree relatives of patients re-porting migraine with aura were 3.8 times more likely to experience migraine with aura than the general population (37). A similar but less robust risk was found in first-degree relatives of those patients reporting migraine without aura. Early twin studies had methodologic flaws (38), but more recent reports show a higher concordance rate among mono-zygotic twins, compared with dizygotic twins, for migraine with and without aura. A large-scale twin survey of the Dan-ish population has estimated that the contribution of genetic effects to susceptibility is 61% for migraine without aura, and 65% for migraine with aura (39). Recent evidence suggests the involvement in some forms of migraine of the calcium channel gene CACNA1A, which is located on chromosome 19p13. Ophoff (40) identified five unrelated families with familial hemiplegic migraine and mutations in CACNA1A. The gene encodes the alpha-1 sub-unit for a brain-specific P/Q type calcium channel. The muta-tions associated with familial hemiplegic migraine, in animal models resulted in the formation of calcium channels with altered biophysical properties (38).A calcium channel defect 1284 CLINICAL NEURO-OPHTHALMOLOGY might lead to recurrent headache through several mecha-nisms, including destabilization of central neuronal control centers involved in the initiation of the attack and disruption of calcium channel functions mediating the release of neuro-transmitters (38,40). Mutations of CACNA1A have also been identified in patients with more common forms of mi-graine. In one family with a history of familial hemiplegic migraine, two individuals who had migraine with aura, but not hemiplegic migraine, had CACNA1A mutations (41). These, and other preliminary data, raise the possibility that migraine is the clinical manifestation of a channel defect. Additional genetic data are needed from other families to determine how widespread CACNA1A mutations are in pa-tients with typical migraine. It is likely that multiple genes are involved in determining susceptibility to, and clinical expression of, migraine. Al-though the genes involved in the serotonergic system are good candidates for susceptibility determination, no associa-tion with migraine with or without aura has been demon-strated for any of the 5-HT receptor subtype genes evaluated (38). The dopaminergic system may also play a role in mi-graine susceptibility, and a correlation between migraine and allele distribution for the dopamine D2 receptor gene has been reported (42). The NcoI allele and the C/C genotype for the D2 receptor gene appear significantly more often in migraineurs than in healthy controls. However, this genotype is also more frequently expressed in individuals with psychi-atric disorders such as anxiety and depression. Since mi-graine frequently co-occurs with depression, this might re-flect a single locus responsible for multiple conditions; alternatively, these findings may represent a clinical artifact. Migraine occurs frequently in patients with mitochondrial disorders, such as mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) (43). Since mi-tochondrial DNA is transmitted maternally, and migraine typically demonstrates maternal inheritance, mitochondrial DNA mutations have been investigated as possible causes of migraine. To date, no association between migraine and known mitochondrialDNAmutations have been established. Future studies will likely identify additional genes in-volved in the initiation and maintenance of migraine. This may in turn help clarify the pathophysiology of the disorder and determine appropriate strategies for prevention and treatment. Migraineand Stroke The relationship between migraine and stroke is complex. Both are neurological disorders associated with focal neuro-logic deficits, alterations in cerebral blood flow, and head-ache. Anumber of case-control and population-based studies have reported an association between ischemic stroke and migraine. The association is more robust for migraine with aura, which appears to be an independent risk factor for stroke (2,44-46). Migrainous infarction is recognized by the IHSas a distinct entity, and describes a cerebral infarction in which the pathogenesis is directly attributable to migraine. Migrainous infarction occurs during a typical migraine with aura, and is exceedingly rare. However, migraines, with and without aura, are associated with ischemic stroke occurring remote in time from a typical migraine attack. Tzourio et al. (45) found a significant association between migraine and ischemic stroke only in women younger than 45 years of age. The association was greater in those patients who smoked. A follow-up study (46) examined female stroke patients younger than 45 years of age, and found an increased risk for ischemic stroke among women with migraine with and without aura. The risk was increased among women who smoked or were using oral contraceptives (OCs). The use of OCs and the risk of stroke in migraineurs has long been a source of controversy. Multiple retrospective studies over the past two decades have evaluated the impact of OCs on cerebral ischemic events in patients with mi-graine; many of these studies were conducted when high-dose estrogen contraceptives were widely used. These data suggest that OCs containing more than 50 mg estrogen are associated with an increased risk of cerebral thrombotic events, while those containing only progestin did not in-crease risk (47). Migrainewith and without Aura Migraine may be divided into two major categories: mi-graine without aura and migraine with aura. Both are clinical syndromes characterized by headache and associated symp-toms. Migraine with aura is also characterized by focal neu-rologic symptoms which typically precede or accompany the headache. Patients with either type may experience a premonitory phase and a resolution phase. These phases may include symptoms such as hyperactivity, hypoactivity, depression, food cravings, and repetitive yawning (48). Migrainewithout Aura Migraine without aura is defined as a recurrent headache disorder manifesting in attacks lasting 4-72 hours. The spe-cific criteria in the IHSsecond edition (11) are: A. At least five attacks fulfilling B to D B. Headache attacks lasting 4-72 hours and occurring less than 15 days/month (untreated or unsuccessfully treated) C. Headache has at least two of the following characteris-tics 1. Unilateral location 2. Pulsating quality 3. Moderate or severe pain intensity 4. Aggravation by, or causing avoidance of, routine physical activity (i.e., walking or climbing stairs) D. During headache, at least one of the following: 1. Nausea and/or vomiting 2. Photophobia and phonophobia E. Not attributable to another disorder When eliciting a history, it should be remembered that many patients have more than one primary headache syn-drome. Migraine without aura is the disease most likely to accelerate in response to the frequent use of symptomatic medications, resulting in medication overuse headache. HEADACHE AND FACIAL PAIN 1285 Therefore, a thorough medication history is an indispensable part of the history. Although there is no mention of age in the above criteria, migraine without aura in children is slightly different. The duration is typically 1-72 hours, and the head-ache is often bilateral; the adult pattern of unilateral head-ache generally develops in late adolescence or early adult life (49). In young children, photophobia and phonophobia may be inferred from behavior. Migraine without aura is the most common type of mi-graine. The average attack frequency is higher than migraine with aura, and the headaches are usually more disabling (2). There is often a strict menstrual relationship. Indeed, pure menstrual migraine and menstrual-related migraine are rec-ognized as distinct subforms of migraine without aura (50). Migrainewith Aura Migraine with aura is a recurrent disorder manifesting in attacks of reversible focal neurologic symptoms, usually developing gradually over 5-20 minutes and lasting for less than 60 minutes. Headache with the features of migraine without aura usually follows the aura symptoms. Less com-monly, the headache may not have migrainous features, or the headache may be completely absent The IHShas de-scribed several subforms of migraine with aura (11); in-cluded here are those most relevant to neuro-ophthalmology. Typical Aura with MigraineHe adache This disorder is characterized by a typical aura consisting of visual and/or sensory and/or speech symptoms. Gradual development, duration of less than 1 hour, and complete reversibility define the aura, which is followed by a headache fulfilling criteria for migraine without aura. The specific IHS criteria (11) are: A. At least two attacks fulfilling criteria B to E. B. Fully reversible visual and/or sensory and/or speech symptoms but no motor weakness. C. At least two of the three following: 1. Homonymous visual symptoms including positive features (flickering lights, spots, lines) and/or nega- Table 26.6 Visual Phenomena Associated with Migraine Visual Event Description Also Caused by Fortification spectra Scintillating scotoma Micropsia Macropsia Metamorphopsia "Alice in Wonderland" syndrome Palinopsia Cerebral polyopia Achromatopsia Arc of jagged, serrated, or zigzag lines Positive fortification spectra on outside and negative scotoma in the middle Objects appear too small Objects appear too large Objects appear distorted in shape Episodes of distorted body image Persistence of visual images Perception of multiple images in both eyes Inability to perceive color Visual seizure, occipital AVM, mass Visual seizure, occipital AVM, mass Visual seizure, occipital AVM, mass, macular disease Visual seizure, occipital AVM, mass Macular disease, visual seizure Visual seizure, occipital AVM, mass Visual seizure; parieto-occipital damage; medications Visual seizure, right occipital lobe damage Bilateral occipito-temporal damage tive features (loss of vision) and/or unilateral sen-sory symptoms. 2. At least one symptom develops gradually within 5 minutes and/or different symptoms occur in succes-sion. 3. Each symptom lasts between 5-60 minutes. D. Headache that meets criteria for B-D for migraine without aura begins during the aura, or follows the aura within 60 minutes. E. Not attributed to another disorder. MigraineAura Visual aura is the most common type of aura, and is gener-ally binocular. The classic description is a fortification spec-tra (2): a zigzag figure that appears near fixation, gradually spreads left or right, and assumes a convex shape with an angulated scintillating edge, leaving variable degrees of ab-solute or relative scotoma in its wake (Fig. 26.7). However, migraineurs may experience a variety of visual phenomena, including complex disorders of visual perception. Table 26.6 summarizes common and uncommon migrainous visual phe-nomena, and the characteristics of the visual aura are detailed in Table 27.7 and Figures 26.8 to 26.12. Metamorphopsia refers to distortion of the shapes of images; the patient may perceive objects as being too fat, too thin, or wavy. Some patients with migraine may describe objects appearing too small (micropsia) or too large (macropsia). Metamorphopsia in general is more commonly associated with macular dis-ease than with migraine. Patients with macular edema fre-quently report micropsia, since the swelling causes increased separation of photoreceptors; macropsia is present with epir-etinal membranes but is rarely a macular symptom. Meta-morphopsia due to macular disease is generally easily distin-guishable from migraine aura: the symptoms are continuous rather than paroxysmal; often unilateral rather than bilateral; and the fundus examination generally shows abnormal mac-ula. Migraineurs may also perceive objects as being too far away (teleopsia). The Alice in Wonderland syndrome typically refers to 1286 CLINICAL NEURO-OPHTHALMOLOGY Figure26.7. Characteristic build-up of a migrainous scintillating scotoma. A small pulsing object is initially noticed in the upper right quadrant of both eyes. The object gradually enlarges to consume the homonymous quadrant, and the border consists of colored, shimmering prisms. The visual disturbance eventually consumes the entire visual field before disintegrating. Figure26.8. Successive maps of a scintillating scotoma showing gradual enlarge-ment of fortification figures and gradual movement away from fixation (x). These drawings were made by Lashley as he observed his own scotomas. (From Lashley KS. Patterns of cerebral integration indicated by the scotomas of migraine. Arch Neurol Psychiatry 1941;46 331-339.) HEADACHE AND FACIAL PAIN 1287 Figure26.9. Two examples of the visual aura of migraine drawn by a 50-year-old patient who had suffered from migraine since adolescence. The visual symptoms began following a prodromal period marked by a feeling of apprehension. She first noted horizontal wavering images, as if every-thing was being observed through a ‘sheet of water.'' A, This drawing depicts the areas of her son's face that were visible and those that were blurred out. The blurred area included the entire right side of the visual field and the inferior left visual field. She stated that: ‘‘Everything looked fluid pinpoint-like globules were present in the area of visual loss . . . it was like snow on a TV screen . . . contours of objects seemed to waver and move like reflections in water . . . there were no zigzags in the periphery . . . it was as if something was wrong with the horizontal adjustment of the TV set . . . all color was gone. The attack lasted about 30 minutes and then I could see again but my face on the right side and my right hand became numb, and then my headache started.'' B, On another occasion, the same patient experienced what she described as a ‘‘sunburst attack.'' It occurred as she was driving down a wooded road. It started as did her usual attacks, and she quickly stopped her car. She remembered that the color green (trees) could be seen only in the central portion of the left field of vision. ‘‘Zigzag bright light surrounded the central area, and tiny molecular droplets danced aimlessly through the image. All forms that could be perceived seemed to wobble like reflections in water. Horizontal ripples passed back and forth across the central field of vision. The periphery of vision was blind and seemed like a silvery haze that pulsated in and out at the zigzag zone.'' This visual aura cleared in 35 minutes and was followed by a right-sided headache and severe nausea. Table 26.7 Characteristics of the Visual Aura in Migraine Characteristic Description Positive or negative phenomena (or both) Visual field Shape Motion Flicker Color Clarity Brightness Evolution/expansion Progression Positive phenomena often occur first, and are followed by negative phenomena Scotoma often start centrally and progress peripherally Scintillations often C shaped; scotoma may be bean shaped Objects may oscillate, rotate, or shimmer Rate usually 10/sec; may change during aura Any; may have no specific color except bright white May be blurry or hazy Often excessively bright Buildup occurs in both scotoma and scintillations Scintillations may march from center to periphery or vice versa Figure26.10. Aerial photograph of an early Italian military fortification demonstrating the angulation often seen in the visual aura of migraine with and without headache. (From Hupp SL, Kline LB, Corbett JJ. Visual distur-bances of migraine. Surv Ophthalmol 1989;33 221-236.) 1288 CLINICAL NEURO-OPHTHALMOLOGY Figure26.11. Hemianopic nature of a fortification scotoma. Note gradual enlargement of the scotoma over 15 minutes. Also note that the scotoma remains localized to the left homonymous hemifield. (From Hupp SL, Kline LB, Corbett JJ. Visual disturbances of migraine. Surv Ophthalmol 1989; 33 221-236.) hallucinations of enlargement, diminution, or distortion of all or part of the body (Fig. 26.13). Lippman (51) originally described seven patients with this syndrome, and recalled that Lewis Carroll (Charles Lutwidge Dodgson), a mi-graineur himself, also described similar hallucinations in his Figure26.12. Visual aura without headache (acephalgic migraine). Pho-tographic representation of a scintillating scotoma that occurred in a patient without headache. Note darkening of the left homonymous visual field associated with blurring of central vision and a fortification figure in the left hemifield. (From Wiley RG. The scintillating scotoma without head-ache. Ann Neurol 1979;11 581-585.) Figure26.13. Drawings of abnormal visual phenomena observed by chil-dren during an attack of migraine. A, Drawing of micropsia as observed by a child suffering from migraine with aura. Other children (right) appeared unusually small to the patient (left) during some of her attacks. B, Drawing by a 9-year-old girl who had episodes that began with double vision. Colors of objects then changed, and a bitemporal headache developed. On one occasion before the headache developed, the child saw a picture as being upside down and her mother as walking upside down. In this drawing, a picture in the room (upper left corner) is upside down, as is the patient's mother (right). (From Hachinski VC, Porchawka J, Steele JC. Visual symp-toms in the migraine syndrome. Neurology 1973;23 570-579.) book, Alice in Wonderland. Similar patients have been re-ported, and the syndrome seems to occur primarily in chil-dren with migraine (52). Although migraine is the cause in most cases of Alice in Wonderland syndrome, other etiolo-gies such as seizure and encephalitis have been reported (53). These distortions of body image are thought to result from cortical depression in the nondominant parietal lobe (54). Migraine may also result in the perception of multiple images: cerebral polyopia. In such cases, the double vision is present with either eye covered, and does not improve with refraction. Palinopsia refers to a persistent visual afterimage: patients may describe seeing a trail of ghost images while watching an object move across the field of vision. Cerebral polyopia and palinopsia have both been associated with le-sions of the nondominant occipital lobe (55,56). Visual allesthesia, or ‘‘upside-down'' vision, may also occur during migraine aura (57). The patient may report vertical or horizontal inversion of the visual environment. HEADACHE AND FACIAL PAIN 1289 Similar phenomena have also been associated with brain-stem lesions involving the vestibular nuclei and central ves-tibular pathways. The time course and the headache follow-ing the visual event suggest a migrainous etiology. Most of the complex visual hallucination and illusions described above can also be caused by visual seizures. This is perhaps not surprising, given the similar phenomenology of migraine and epilepsy. Similar phenomena have been re-ported to occur in patients with irritative occipital mass le-sions, particularly vascular malformations (57). Distinguish-ing a visual seizure or focal occipital lesion from migraine aura may in most cases be accomplished through a complete history and examination, which must include perimetry. The gradual onset and evolution of symptoms common to mi-graine aura is not seen with a visual seizure, which is maxi-mal at onset. Visual seizures are generally shorter than mi-graine auras, lasting 3-5 minutes. In many patients with visual seizures, a visual field defect attributable to a fixed posterior visual pathway lesion may be identified. Alcohol and illicit drugs may cause visual hallucinations similar to migraine aura. Certain prescription and over-the-counter medications may cause visual phenomena as well; a thorough medication history and a temporal relationship between drug ingestion and visual symptoms suggests the etiology (57). Regardless of the complexity, a migraine aura should still fulfill the criteria for aura described above; any variation (i.e., rapid onset without gradual evolution, duration too brief or too prolonged) should raise suspicion about an alternate etiology. It bears repeating that complete reversibility is the hallmark of an aura; a focal deficit after the aura has resolved mandates further evaluation, including neuroimaging. Typical Aura without Headache This disorder has carried several names over the years, including ‘‘acephalgic migraine.'' The aura fulfills all the criteria detailed above. Patients with typical aura without headache commonly have a history of migraine with aura (2). As patients with typical aura age, the headache may lose migraine characteristics or disappear completely even as auras continue. Some patients, primarily males, will have migraine aura without headache exclusively (58). The differ-entiation between migraine aura and transient ischemic at-tack may be difficult in the elderly, but can generally be accomplished with a thorough history and examination. In the Framingham cohort, the incidence of migrainous visual symptoms was 1.33% for women and 1.08% for men. These episodes often occurred after 50 years of age, independent of concurrent headaches and, in some cases, in the absence of a history of recurrent headaches. These episodes did not appear to be associated with an increased incidence of stroke (59). Fisher described ‘‘later life migrainous accompani-ments,'' or migraine aura (often not accompanied by head-ache) occurring in older patients, some with a previous his-tory of migraine headache (60). He identified a number of clinical features suggestive of migraine accompaniment, in-cluding binocular visual symptoms, build-up of scintilla- Table 26.8 Later-Life Migrainous Accompaniments The gradual appearance of focal neurologic symptoms, spreading or intensifying over minutes Positive visual symptoms characteristic of typical migraine aura: Scintillating scotoma, flashing lights, shimmering Duration of 15-25 minutes History (may be remote) of similar symptoms associated with migrainous headache The occurrence of two or more identical spells "Flurry" of accompaniments (often in the 50-60-year age group) Generally benign course (Adapted from Fisher CM. Late-life migraine accompaniments as a cause of unexplained transient ischemic attacks. Can J Neurol Sci 1980;7 : 9-17.) tions, and duration of 15-20 minutes (Table 26.8). He also reported a midlife ‘‘flurry'' of identical episodes, and a be-nign course. Therefore, in patients with typical presentations (binocular symptoms, gradual evolution, duration less than 1 hour) and a normal neurologic and neuro-ophthalmic ex-amination (including visual field testing), extensive diagnos-tic testing is not routinely indicated. In patients with atypical presentations (brief duration, lack of build-up), noninvasive testing (MR imaging and angiography, echocardiography) is reasonable. Retinal Migraine Retinal migraine is defined by repeated attacks of mono-cular visual phenomena including scintillations, scotomata, or blindness associated with typical migraine headache (2). The aura and migraine headache fulfill all the criteria de-scribed earlier, with the only distinction being the monocular visual symptoms. Again, history is paramount. Since the visual field is slightly larger temporally than nasally, patients with homonymous visual field defects often report monocu-lar visual loss, localizing it to the eye with the temporal field cut. The patient should be asked specifically whether each eye was alternately covered during the attack. Retinal mi-graine without headache may occur, but is rare (61). Retinal migraine may be mistaken for an episode of amaurosis fugax, i.e., transient monocular visual loss due to a thrombo-embolic transient ischemic attack and should only be consid-ered after all other causes of transient monocular loss have been ruled out. The presence of positive visual phenomena is much more suggestive of migraine, although such symptoms have been reported in carotid occlusive disease (62). Dura-tion of 1-10 minutes, sudden rather than gradual onset of symptoms, and lack of typical migraine headache following visual loss would all suggest a thrombo-embolic rather than migrainous etiology. Persistent Migraine Aura without Infarction Although migraine aura by IHScriteria lasts less than 60 minutes, persistence of aura for days or years may rarely occur (63), and is included in the second edition of IHS criteria (11). The IHScriteria require previous attacks fulfill- 1290 CLINICAL NEURO-OPHTHALMOLOGY ing criteria for migraine with aura, similarity of aura symp-toms to previous attacks, and exclusion of alternate etiolo-gies, including (as the name implies) migrainous infarction. Indeed, prolonged, continuous photopsias are more sugges-tive of retinal disease, with involvement of the deep retinal layers. Patients with severe, bilateral visual loss may experi-ence ‘‘release'' hallucinations; these are typically formed (57). Certain medications, most commonly psychoactive drugs, may be associated with positive visual phenomena as well. Occipital seizures typically manifest as paroxysmal rather than continuous photopsias; these photopsias are gen-erally colored, whereas migraine aura is typically achromatic (64). Persistent occipital seizures (‘‘status epilepticus amau-roticus'') are rare, but have been reported (65). Therefore, patients presenting with continuous photopsias require a careful headache and medication history, a complete oph-thalmic examination (including visual fields and evaluation of the peripheral retina), and in some cases, ancillary testing (such as neuroimaging, EEG, ERG). Treatment of persistent migraine aura is notoriously unsatisfactory, but medications that have been used include amitriptyline and gabapentin. Familial Hemiplegic Migraine The IHSdefines familial hemiplegic migraine (FHM) as recurrent migraine with aura including motor weakness in which at least one first- or second-degree relative has mi-graine aura including motor weakness. The disorder has au-tosomal dominant inheritance, with variable penetrance. The IHScriteria (11) include: A. At least two attacks fulfilling B-E. B. Fully reversible motor weakness and at least one of the following other aura symptoms: visual, speech, or sensory disturbance. C. At least two of the following: 1. At least one aura symptom develops gradually over more than 5minutes and/or different symptoms occur in succession. 2. Each aura symptom lasts less than 24 hours. 3. Headache that meets criteria for migraine without aura begins during aura or follows aura within 60 minutes. D. At least one first- or second-degree relative has mi-graine attacks with aura, including motor weakness. E. Not attributed to another disorder. Onset is typically in childhood, and nearly always by the age of 30 years. Onset of FHM after the age of 40 would be extremely atypical, and should prompt consideration of other etiologies. Many patients with FHM may also experi-ence typical migraine with or without aura (2,66). As noted previously, specific genes have been associated with FHM. Familial hemiplegic migraine type 1 (FHM1) localizes to chromosome 19, and familial hemiplegic mi-graine type 2 (FHM2) is localized to chromosome 1. Patients with FHM type 1 often have basilar type symptoms in addi-tion to the typical aura, and may experience disturbance of consciousness, fever, CSF pleocytosis, and confusion during an attack. In most patients the attacks decline in frequency and terminate at between 30-40 years of age. The neurologic signs are always completely reversible, even in patients with prolonged attacks. Brain MRI typically remains normal. Treatment of FHM differs somewhat from other forms of migraine. Medications with vasoconstrictor effects are best avoided. Severe and persistent hemiplegic migraine has been treated with intravenous dopamine agonists, calcium channel blockers, and acetazolamide (66). Basilar TypeMigraine Basilar type migraine describes a recurrent attack of mi-graine with aura in which symptoms clearly originate from the brainstem or both hemispheres, and where no weakness is present. A headache that meets criteria for migraine with-out aura begins during the aura or follows aura within 60 minutes. This type of migraine is rare and should only be diagnosed after vertebrobasilar ischemia has been ruled out. Ophthalmoplegic Migraine Ophthalmoplegic migraine is now classified as a cranial neurelgia by the IHSand is discussed later. Pathophysiology of Migraine Migraine may be considered as a sensitivity of neurovas-cular reactions to certain stimuli or cyclic changes in the central nervous system (2). Any theory of migraine patho-genesis must encompass the initiation, amplification, and maintenance of head pain, as well as the neuroanatomic and neurophysiologic substrates of aura. Traditional theory of migraine pathogenesis has included a vasogenic etiology and a neurogenic etiology. The vasogenic theory is derived in part from observations made by Wolff and colleagues in the 1930s. Graham and Wolff (67) recorded the pulsations of branches of the superficial temporal artery during migraine headache and observed that the amplitude of the pulse wave diminished as the intensity of the headache was reduced after the injection of ergotamine, a vasoconstrictor. Further, stimulation of intracranial vessels in awake patients during craniotomy caused severe, ipsilateral, throbbing headache. It was felt that intracranial vasoconstriction was responsible for the aura, and the rebound dilation and distension of extra-cranial vessels resulted in headache. The neurogenic theory holds that the brain is the generator of migraine, and that individual susceptibility reflects differential thresholds. In this model, the vascular changes observed during the head-ache are the effect rather than the cause. In the past two decades, it has become clear that neither of the traditional theories is adequate to explain all the phenomena that accom-pany migraine. Over the past decades, advances in neuroim-aging have elucidated the mechanism of aura and given insight into potential molecular mediators of headache trans-duction and generation. In particular, functional imaging modalities, such as positron emission tomography (PET) and functional MRI (fMRI), allow direct observation of physio-logic rather than anatomic abnormalities. Leao (68) originally described a ‘‘spreading depression'' or progressive shutdown in cortical function in animal brain, and felt that this may be related to migraine aura. Cortical HEADACHE AND FACIAL PAIN 1291 stimulation initiates excitation followed by depression of neuronal activity that spreads slowly from the focal site of stimulation at rates ranging from 2-6 mm per minute. Al-though pial arterial and venous dilation occur simultaneously with the neural activity, spreading depression does not fol-low vascular boundaries. Functional neuroimaging has al-lowed detailed investigation of cortical activity during mi-graine aura. Olesen (69) monitored regional cerebral blood flow (rCBF) in patients with aura-like symptoms undergoing carotid angiography and found a 25-30% decrease in rCBF which began in the occipito-temporal region and spread ante-riorly at a rate of 2-3 mm per minute without respecting vascular boundaries. This spreading oligemia has been ob-served by other investigators using a variety of neuroimaging techniques (70). The mechanism by which aura transduces headache re-mains unclear. The trigeminal nucleus caudalis, which is felt to be part of the central pathway mediating migraine pain, is stimulated by spreading depression. Many of the trigemi-novascular pathways discussed earlier are likely to be in-volved in the amplification and maintenance of migraine head pain. Sterile neurogenic inflammation and plasma pro-tein extravasation occurs in response to activation of trigemi-nal sensory neurons (70,71). This results in meningeal in-flammation that persists for minutes to hours. Although neurogenic inflammation is well demonstrated in animals, it is not clear whether this response occurs in humans with migraine. Gene regulation studies in mice have suggested that atrial natriuretic peptide in glia may be an important mediator (72). There may also be molecules that diffuse through the CSF and stimulate the pial vessels that are yet to be identified. The mechanism by which migraine headache is generated without aura is uncertain. However, several inves-tigators have demonstrated cortical events similar to spread-ing depression in patients with migraine without aura. Woods (73) found bilateral spreading oligemia beginning occipitally in a patient with migraine but no aura. One study using fMRI found spreading depression in the occipital cor-tex before headache onset in migraine patients with and with-out aura. This implies that similar primary neuronal and sec-ondary hemodynamic events may precede initiation of headache in all patients. Another mechanism may be the recruitment of cortical-subcortical connections to nocicep-tive centers via the wave of excitation-suppression invading the cortical components of these networks (70). In this fash-ion, aura without headache may be an example in which the spreading depression does not involve the cortical-subcorti-cal connections to brainstem nociceptive networks (70,74). The observation that migraine attacks originate in the brain and can be triggered under certain condition implies the existence of a threshold that governs the incidence of attacks. There is a prevailing theory that migraineurs have transient or persistent hyperexcitability of neurons in the cerebral cortex, particularly the occipital lobe. Although some studies have yielded conflicting results, data have con-sistently supported cortical hyperexcitability (75,76). Trans-magnetic stimulation of the occipital cortex required to pro-duce visual phosphenes is lower in migraine patients with aura than in controls (77). Battelli (78) reported a lower phosphene threshold for transmagnetic stimulation over Brodmann's area V5 in migraineurs, suggesting that hyper-excitability extended beyond the primary visual cortex (V1). Further studies, using magnetoencephalography and fMRI, have confirmed widespread regions of abnormal cortical ex-citability in patients with migraine (76). Cell membrane ex-citability appears to be the critical determinant for suscepti-bility to migraine attacks. There may be multiple factors that increase or decrease neuronal excitability, and modulate the attack threshold. There is increasing evidence suggesting an important role of serotonin in migraine. In the 1970s, the serotonin receptor mediating constriction of blood vessels was characterized, and shown to correspond to a subtype labeled ‘‘5-HT1- like.'' This led to the development of sumatriptan, a specific agonist of the 5-HT1b, 5-HT1D, and 5-HT1F receptors. Since that time, a large number of serotonin agonists (the triptans) have entered the market, and have revolutioned the abortive treatment of migraine attacks; 5-HT1B/1D receptor agonists block electrical activity in the trigeminocervical re-gion after stimulation of pain-sensitive intracranial struc-tures (2,70). It is uncertain whether these agonists act periph-erally, centrally, or both. The blood-brain barrier is relatively impermeable to sumatriptan, a 5-HT1B/1D agonist, which only exerts an inhibitory effect on the trigeminocervical nu-cleus when the blood-brain barrier has been disrupted. How-ever, there is indirect evidence that the blood-brain barrier may be altered during migraine (2,79). Hoskins and Goadsby (80) have speculated that the antimigraine efficacy of trip-tans is due to a direct inhibitory effect on second-order tri-geminal neurons. 5-HT1F receptors are located on second-order trigeminal neurons, and not on vascular smooth muscle cells. Animal models indicate that selective 5-HT1F agonists might be effective in the treatment of migraine, without the detrimental cardiovascular side effects of the 5-HT1B/1D ago-nists (76). This has not yet been replicated in human trials. Rostral brainstem structures, such as the periaqueductal gray and locus ceruleus, may also be involved in migraine genera-tion and maintenance (70,76). A theory of central neuronal hyperexcitability has been proposed, in which common pathways mediating head pain are more easily triggered in patients with episodic migraine. It is clear that migraine pain is the result of a complex inter-play between peripheral activation of nociceptors and central modulation. Migraineurs carry a predisposition to activation of these pathways, most likely explainable on a cellular and genetic basis. As functional neuroimaging continues to evolve, and molecular and genetic mediators are more pre-cisely defined, the mechanisms underlying migraine may be further elucidated. Treatment of Migraine Effective management of migraine headaches begins with making an accurate diagnosis. Despite the fact that migraine is treatable, many patients find themselves in a paradoxic situation in which they are both undertreated and overmedi-cated: they take or receive a large number of nonspecific analgesics which are minimally effective for migraine and contribute to analgesic rebound. Treatment of migraine should ideally include nonpharmacologic as well as pharma-cologic therapy. Minimizing excessive stress and sleep de- 1292 CLINICAL NEURO-OPHTHALMOLOGY privation, eating regular meals, exercise, and avoidance of common triggers (e.g., wine, chocolate, etc.) are practical measures that should be discussed with any migraine patient. Alternative therapies (biofeedback, relaxation, acupuncture) have been used successfully in conjunction with conven-tional medical treatment (81). In many patients, a headache diary is invaluable; as it may help the patient and the physi-cian characterize the headaches more clearly. A headache diary may also give patients a sense of control over their illness. There is an ever-expanding armamentarium of medica-tions available for the treatment of migraine, which can be daunting to practitioners. Treatment of migraine may be di-vided into acute or abortive treatment, and preventive or prophylactic treatment. Abortive Treatments A complete review of all abortive treatments for migraine is beyond the scope of this chapter. A recent review (82) graded the quality of evidence for medications used for acute migraine treatment. The following medications had grade A recommendations, meaning that they were based on multi-ple, well-designed clinical trials. Nonsteroidal anti-inflammatory drugs (NSAIDs) are appropriate agents for patients with mild to moderate mi-graine pain. Specific agents include acetaminophen, ibupro-fen, naproxen sodium, and aspirin. Their effect is better when the drug is taken at the beginning of the migraine attack, ideally during the prodromes. All triptans are serotonin receptor agonists with approxi-mately equal affinity for the 5-HT1B/1D receptor. They all inhibit the release of vasoactive peptides, promote vasocon-striction, and block pain pathways in the brainstem (70). They are effective and relatively safe for acute therapy of migraine, and are useful choices for patients with moderate to severe migraine. Routes of administration differ among the agents, and intranasal and subcutaneous forms may be beneficial for patients who cannot take oral drugs (secondary to nausea). Although there are pharmacologic differences among the agents, head-to-head efficacy comparison studies are lacking, and there are no clear guidelines in choosing one agent over the other (82). All of the available triptans are contraindicated in patients with cardiovascular disease. Unlike other treatments of acute migraine attack, they may be taken at any time during the migraine; some authors avoid their administration during the aura because of their vaso-constrictive properties. Ergot alkaloids and derivatives include dihydroergo-tamine (DHE) and ergotamine as members of the ergot fam-ily used in migraine treatment. DHE is recommended for patients with moderate to severe migraine pain. DHE may be delivered subcutaneously, intravenously, intramuscularly, or intranasally. Oral or rectal ergotamine may be useful for certain patients with moderate to severe migraine, but the evidence for efficacy is inconsistent, and side effects are more frequent compared with other abortive agents. They should be taken at the beginning of the migraine attack. Opioid analgesics must be balanced against the risks of sedation and dependency. Oral opioids may be useful when sedation will not place the patient at risk, and the potential for abuse is discussed in advance. Intranasal butorphanol, a mixed opiate receptor agonist-antagonist, may be useful for some patients, but carries the same risk of dependency. Antiemetics, are often used as adjuncts to other migraine therapies, but certain drugs in this class may have indepen-dent pain-relieving effects. Intravenous metoclopramide and intravenous and intramuscular prochlorperazine may be con-sidered as monotherapy for migraine pain. Oral antiemetics are suggested only for adjunctive use. Preventive Treatments of Migraine The main goal of preventive treatment of migraine is to reduce the frequency, severity, and duration of subsequent migraine attacks. Preventive treatment plans often allow for the use of agents for acute attacks. In general, preventive treatment is considered when the frequency of headache is so high that analgesic rebound is likely, when migraine has a substantial impact on a patient's life despite the use of abortive agents, and when contraindications may limit the use and success of acute medications (2). Nonpharmacologic treatments, as discussed above, should be incorporated into the overall management plan. The relative efficacy of phar-macologic versus nonpharmacologic treatment is undeter-mined (83). In the United States, only methysergide, propranolol, ti-molol, and divalproex sodium are approved for the preven-tive treatment of migraine. However, several other medica-tions have been used, with varying success. The mechanisms of action of the preventive medications are varied, and no single medication had emerged as a clearly dominant treat-ment. The choice of drugs depends upon the goals of the patient, and may be influenced by patient lifestyle issues, medical co-morbidities, side effect profile, and cost of the drug. The U.S. Headache Consortium has classified migraine prophylactic medications in five groups based on clinical efficacy, adverse effects, safety profile, and expert clinical experience (83). Group 1 drugs have high proven clin-ical efficacy in migraine treatment, with mild to moderate adverse effects. The use of the medications is supported by Class A scientific evidence, i.e., their efficacy has been shown in multiple, well-designed, randomized trials. Group 1 drugs include the anticonvulsant divalproex sodium, the tricyclic antidepressant amitriptyline, and the beta-blockers propranolol and timolol. These are considered first-line treat-ment in the prophylaxis of migraine. Group 2 drugs have lower efficacy than group 1, and mild to moderate side ef-fects. Atenolol, verapamil, aspirin, naproxen, gabapentin, fluoxetine, and magnesium are considered group 2 drugs. Group 3 drugs have efficacy based only on clinical experi-ence, but no scientific evidence. This group includes the tricyclic antidepressants nortriptyline and protriptyline, the serotonin-specific reuptake inhibitors sertraline and paroxet-ine, and the anticonvulsant topiramate. Group 4 drugs have medium to high efficacy, but potentially significant side ef-fects. Flunarizine and methysergide are group 4 drugs. Group 5 drugs have no evidence showing efficacy over HEADACHE AND FACIAL PAIN 1293 Table 26.9 Preventive Migraine Medications with Group 1 or Group 2 Level Efficacy Medication Class Dose (mg/day) Side Effects Avoid in Patients with Consider in Patients with Divalproex sodium Propranolol Amitriptyline Gabapentin Verapamil Anticonvulsant Beta-blocker Tricyclic antidepressant Anticonvulsant Calcium channel blocker 500-1500 80-240 50-150 900-2400 240 Weight gain, tremor, alopecia, elevated liver enzymes Weight gain, depression, syncope Dry eyes and mouth, drowsiness Drowsiness, dizziness, weight gain Hypotension, constipation Hepatic disease Asthma, diabetes mellitus, depression Cardiac arrhythmia, elderly patients No specific contraindications Orthostatic hypotension Bipolar disorder, seizure disorder Cardiovascular disease, hypertension Depression, tension-type or chronic daily headache Seizure disorder, neuropathic pain Cardiovascular disease, migraine aura placebo. This group includes clonidine, carbamazepine, and indomethacin. Table 26.9 shows the common preventive agents used for migraine. In recent years, botulinum toxin has been reported to be effective in the treatment of migraine (84). The treatment protocol generally involves injections of botulinum toxin into the head, facial, and cervical musculature. The pre-sumed mechanism involves decreasing or eliminating pe-ripheral activation, and thereby reducing central trigemino-vascular activation (85). However, to date, there have been no well-designed, randomized, blinded and controlled trials demonstrating efficacy. Further, in many reports, the specific headache types have been inadequately characterized. Therefore, there in no Class A evidence favoring the routine use of botulinum toxin in patients with migraine. TENSION-TYPE HEADACHE Tension-type headache (TTH) is the most common form of primary headache, with a lifetime prevalence ranging from 30-78%. However, it remains the least studied of the primary headache syndromes, despite the fact that TTH may have the highest socioeconomic impact (86). Further, mi-graine headache is often misdiagnosed as TTH. In one study, 37% of patients originally diagnosed with TTH were later found to have migraine (87). The situation is complicated by the fact that migraine and TTH may coexist in the same patient. Indeed, the symptoms of each headache type may overlap. Muscle tension is often considered a unique feature of TTH, while migraine is commonly felt not to be associated with muscle tension or neck pain. Recent studies have sug-gested that neither of these assumptions may be correct (88). One study demonstrated that of 144 patients meeting IHS criteria for migraine, 75% described neck pain associated with the attacks (89). The second edition of the IHScriteria (11) divides TTH into episodic and chronic forms. Episodic TTH is further subdivided into frequent and infrequent forms. Episodic TTH Episodic TTH is characterized by frequent episodes of headache lasting minutes to days. Patients often describe the pain as pressing or tightening in quality. The head pain is usually bilateral, of mild to moderate intensity, and does not worsen with routine physical activity. Nausea is absent, but photophobia or phonophobia (but not both) may be present. The headaches occur fewer than 15 days a month for at least 3 months. Episodic TTH may be seen in patients with coexisting migraine without aura, and it may be difficult to distinguish between the two. In one population-based study, 83% of individuals with headaches meeting IHScriteria for migraine also had headaches meeting IHScriteria for ten-sion- type headache (90). A headache diary is an important tool to help differentiate between the two headaches, and guide treatment. Chronic TTH Chronic TTH is described as daily or very frequent head-aches lasting minutes to days. The characteristics of the headache are identical to episodic TTH, but the attacks occur more than 15 days per month for at least 3 months. The patient may have no more than one of the following: photo-phobia, phonophobia, or mild nausea. The patient must be using analgesics less than 10 days per month; if not, the headache is classified as medication overuse headache. Pathophysiology of Tension-Type Headache The exact mechanism of TTH remains undetermined. The traditional theory was that TTH developed due to sustained contraction of pericranial muscles (2). However, muscle contraction may occur to the same degree in migraineurs, and there is no correlation between muscle contraction, ten-derness, and the presence of headache (91). It is likely that peripheral pain mechanisms act as a trigger for a central process. Peripheral mechanisms may be most important in episodic TTH, while in chronic TTH, central mechanisms may predominate. Some have argued that TTH and migraine may represent different portions of a distribution of painful, episodic headaches (2, 92). It is not clear whether episodic TTH may be a mild form of migraine without aura. 1294 CLINICAL NEURO-OPHTHALMOLOGY Treatment of TTH There are few controlled studies guiding the treatment of TTH. Acute therapy typically involves the structured use of simple analgesics (acetaminophen, aspirin), alone or in combination with NSAIDs. Overuse of these drugs must be limited. Prophylactic therapy should be considered when the frequency, duration, and severity of the headache might lead to disability or the overuse of acute medications. The tri-cyclic antidepressants are the drugs most commonly used, although with little scientific evidence. Serotonin-specific reuptake agents (SSRIs) may be used as well. Nonpharmaco-logic treatment, such as biofeedback and relaxation therapy, may be beneficial (2). TRIGEMINAL AUTONOMIC CEPHALGIAS This group of headache syndromes shares the clinical fea-tures of head and facial pain with prominent cranial auto-nomic features (11). Although relatively rare, these disorders are important to recognize since they often display an excel-lent and highly selective response to treatment. In many of these syndromes, the pain is localized to the eye, so the ophthalmologist may be the first health care provider to evaluate the patient, or may see the patient in consultation for unexplained eye pain. Based upon their common clinical phenotype, these headaches have recently been grouped under the term Trigeminal Autonomic Cephalgias (TACs) (11,93). All of these headache types may be associated with structural and vascular lesions, which should be excluded by a combination of careful history and examination, and judiciously chosen ancillary studies. Pathophysiology of Trigeminal Autonomic Cephalgias: Overview It is likely that all of the TACs share a similar pathophysi-ology involving a disturbance in the trigeminal-autonomic reflex. Stimulation of the trigeminal ganglion in animals re-sults in increased extracerebral and cerebral blood flow (93). The major portion of the vasodilator response is mediated via a reflex connection with the facial nerve, the cranial parasympathetic outflow. The parasympathetic arm of this reflex is active in cluster headache (94). Vasoactive intes-tinal peptide (VIP) and CGRP may be important mediators in this response. Some investigators have suggested that cluster headache originates from the cavernous sinus, based on the intersection of the first division of the trigeminal nerve and the cranial sympathetic and parasympathetic nerves at this location. In one case, cerebral angiography demonstrated narrowing of the cavernous carotid artery during cluster (95). However, some authors have suggested that extreme oph-thalmic division pain of many types can result in flow changes in the cavernous sinus, as part of the trigeminal sympathetic reflex (93, 96). Some TACs, such as cluster headache, demonstrate seasonal cycles, implicating the su-prachiasmatic nucleus (the so-called hypothalamic pace-maker) and higher, diencephalic centers. Indeed, PET scans obtained in patients with cluster demonstrate unilateral acti-vation in the hypothalamic gray (93). Further investigation, including the use of functional neuroimaging, may yield more insight into the processes driving TACs. Cluster Headache Cluster headache has carried a number of names over time, including Horton's headache, petrosal neuralgia, cili-ary neuralgia, and others (2). Cluster headache is character-ized by recurrent episodes of severe, unilateral orbital or temporal pain. Most patients are restless and agitated during the attack, as opposed to migraine patients, who often prefer to be still. Men are afflicted 3-4 times more frequently than woman. The IHSrecognizes both an episodic and chronic cluster headache (11). The chronic form may arise de novo, or may evolve from the episodic form. In the episodic form, the patient must have an attack-free interval of one month. Most individuals with episodic cluster headache experience one or two attack phases per year. Phase duration may range from 4-16 weeks. Each attack phase is followed by a remis-sion period of 6 months to 2 years (2,99). Patients with chronic cluster headache experience attacks for greater than 1 year, with remission periods lasting less than 1 month. The IHScriteria for cluster headache are (11): A. At least five attacks fulfilling criteria B to D. B. Severe or very severe unilateral orbital, supraorbital, and/or temporal pain lasting 15-180 minutes, untreated for more than half the period. C. Headache is accompanied by at least one of the follow-ing symptoms or signs, which must be ipsilateral to the pain: 1. Conjunctival injection. 2. Nasal congestion, or rhinorrhea, or both. 3. Eyelid edema. 4. Forehead and facial sweating. 5. Miosis, ptosis, or both. 6. Headache is associated with a sense of restlessness of agitation. D. Frequency of attacks: From one every other day to 8 per day for more than half the period or time if chronic. E. Not attributed to another disorder. The pain is typically excruciating, and has in a few refrac-tory cases resulted in suicide. The attacks are often reliably provoked by alcohol and nitroglycerin. Cluster headache may occasionally be associated with tri-geminal neuralgia: the Cluster-Tic syndrome. Both condi-tions must be treated for the patient to be headache free. It is important to exclude disorders that mimic cluster headache, including other TACs, carotid artery dissection, pituitary ad-enoma, cavernous carotid and anterior communicating artery aneurysms, and Tolosa-Hunt syndrome (2,97). The patient's age, gender, the characteristics of the attacks, and the clinical examination usually serve to differentiate secondary causes of cluster-like headaches from primary cluster. For example, a middle-aged man with a long history of paroxysmal attacks fulfilling critertia for cluster headache and a normal interictal examination is unlikely to harbor a secondary cause of head-ache, and the diagnosis can rest upon clinical criteria. Ap- HEADACHE AND FACIAL PAIN 1295 proximately one-half to two-thirds of patients will develop Horner's syndrome during the attack; after multiple attacks, a permanent Horner's syndrome may result. The mechanism of Horner's syndrome may result from dilation of the inter-nal carotid artery and compression of the sympathetic plexus. (97). Alternatively, dilation and vascular congestion within the cavernous sinus may involve the sympathetic plexus and lead to a transient Horner's syndrome. Pathophysiology of Cluster Headache The pathophysiology of cluster headache is unknown. There is growing evidence that the pain and autonomic symptoms result from dual activation of the trigeminal vas-cular and cranial parasympathetic pathways. Goadsby (94) demonstrated increased levels of CGRP (a marker peptide for the trigeminal vascular system) and VIP (a marker for the cranial parasympathetic system) in the internal jugular vein drainage during attacks. The CGRP levels returned to normal after successful treatment and resolution of the pain. There is evidence that nitric oxide (NO) may play a role in cluster, as well in other TACs and perhaps migraine. NO is increased in patients with cluster headache during and be-tween attacks, and nitroglycerin can consistently provoke attacks in patients with cluster headache (97). These observa-tions, however, do not address the essential feature of the disorder: periodicity. Hypothalamic dysfunction has long been postulated to be a potential cause of the rhythmic nature of cluster, and several studies have reported altered circadian rhythms of hypophysial hormone systems in cluster (98). The suprachiasmatic nucleus (the so-called hypothalamic pacemaker) is regulated by photic stimuli, and governs the synthesis and secretion of melatonin from the pineal gland. A small study demonstrated efficacy of melatonin in cluster headache (99), although the response rate was low. May and colleagues (100) used PET to evaluate variations in regional cerebral blood in nine patients during a cluster attack. These scans were compared with eight patients with chronic cluster headaches who were not in the midst of a cycle. The results suggested that the medial hypothalamus is specifically in-volved in cluster headache; hypothalamic activation did not occur when nitroglycerin was given outside of an attack, or with experimentally induced first trigeminal division pain. Hypothalamic activation may therefore act in a triggering or permissive manner. Therefore, dual activation of the trigeminal vascular and cranial parasympathetic system occurs during a ‘‘cluster-permissive'' period. This period is determined and modu-lated by dysfunction in the hypothalamus. Activation of the two systems results in the release of marker peptides, which leads to vasodilation, vascular congestion, and possibly neu-rogenic inflammation. This cascade leads to pain and auto-nomic symptoms (97). Treatment of Cluster Headache As with migraine, treatment of cluster may be either acute or preventive. A limiting factor for acute treatment is that the attacks reach peak intensity very quickly, and are of relatively short duration. Therefore, oral preparations used for migraine attacks may not have a rapid enough onset. Inhalation of 100% oxygen for 15 minutes can arrest an attack (97). The injectable forms of sumatriptan and dihy-droergotamine have been shown to be extremely effective for acute treatment of cluster headache (101). Oral triptan preparations have shown modest to no benefit (102). Intra-nasal lidocaine may be of benefit (2). Prophylactic treatment is used to shorten attacks and re-duce attack frequency. Preventive therapy is frequently re-quired for cluster since the attacks are frequent, severe, and often too short-lived for abortive medications to take effect. Also, overuse of abortive treatments may result in analgesic rebound. A typical regimen involves starting the medication early in the cluster period, continuing the drug until the pa-tient is headache free for 2 weeks, and then tapering the medication. Prednisone is often used as a short-term, transi-tional therapy, and is the most rapid-acting of the preventive drugs used for cluster headache (101). A typical protocol might be 2-3 weeks of prednisone in tapering doses to break the cycle of headache until other medications become effec-tive. However, steroid-dependence is often seen in these pa-tients who may have rebound cluster headaches. Verapamil, a calcium channel blocker, has proven to be effective for cluster prophylaxis, and is considered by many to be the first-line drug in preventive therapy. Ergotamine tartrate may quickly suppress attacks, but is used less frequently, due to possible physiologic dependence and a contraindication to the concurrent use of sumatriptan for acute treatment. Lith-ium is highly effective to prevent cluster headaches, and is well tolerated when monitored carefully. Methysergide may also be used for maintenance therapy, but the possible retro-peritoneal fibrosis associated with methysergide limit rou-tine use of this medication. Valproic acid, gabapentin, and topiramate are anticonvul-sants that, in open label trials, have been found to be useful for cluster prevention (97). In one open label study, valproic acid produced a response in 73% of subjects with cluster headache (103). Refractory cluster may occur in a small number of patients. In these cases, it may be worthwhile to reevaluate for secondary causes of cluster, such as a com-pressive or vascular lesion. Combination preventive therapy may occasionally be required, and selected patients may ben-efit from a short course of intravenous DHE or methylpred-nisolone. Ablative surgical procedures may be considered in patients who are truly medically refractory, either due to ineffectiveness of medication, or contraindications, or intol-erable side effects. The most frequent procedure is radiofre-quency trigeminal rhizotomy (97). In a prospective study of 17 patients with chronic, intractable, cluster headache who underwent complete or partial trigeminal root resection, 88% had complete or nearly complete relief of their symptoms after 5 years of follow-up (104). Future surgical treatments may attempt to specifically target the anatomic structures involved in the pathogenesis of cluster. A single case study described the stereotactic implantation of a stimulating elec-trode into the homolateral posterior hypothalamus of a pa-tient with medically refractory chronic cluster headache. This resulted in complete remission of headache after 6 months of follow up (105). Controlled, randomized trials 1296 CLINICAL NEURO-OPHTHALMOLOGY will be necessary to determine whether these procedures will become part of a routine treatment algorithm for cluster pa-tients. Chronic Paroxysmal Hemicrania Chronic paroxysmal hemicrania (CPH) was first de-scribed in 1974 (93), and is characterized by frequent, short-lasting attacks of unilateral pain, typically in the orbital, supra-orbital, or temporal region. The IHScriteria are (11) as follows: A. At least 20 attacks fulfilling B to E. B. Attacks of severe, unilateral orbital, supraorbital, or temporal pain lasting 2-30 minutes. C. Attack frequency greater than five per day for more than half the time. D. Pain is associated with at least one of the following symptoms or signs ipsilateral to the pain: 1. Conjunctival injection, lacrimation, or both. 2. Nasal congestion, or rhinorrhea, or both. 3. Eyelid edema. 4. Forehead and facial sweating. 5. Miosis, ptosis, or both. E. Headache is stopped completely by indomethacin. F. Not attributable to another disorder. An episodic form of paroxysmal hemicrania (EPH) has been reported, but is less common; EPH occurs in periods lasting from 7 days to 1 year, and separated by remissions of at least one month. In CPH, the attacks occur for more than 1 year without remission, or with remissions lasting less than 1 month (2,93). The episodic and chronic forms occur more frequently in women (3 1), with an age range of 6-81 years (mean 34 years). The pain is excruciating, may be throbbing or boring, and builds up rapidly. In contrast to cluster, patients with CPH prefer to be still. CPH and cluster share many character-istics, which may lead to diagnostic uncertainty. Indeed, the final common pathways (activation of trigeminal vascular and cranial parasympathetic outflow systems) are likely sim-ilar in both disorders. As with other headache syndromes, secondary causes of CPH should be considered and excluded through history, examination, and selected studies. Second-ary CPH has been associated with structural lesions of the frontal lobe and sella turcica (93), cavernous sinus meningi-oma (106), and collagen vascular disease (107). As with cluster headache, autonomic features are a defining feature of the disorder, and their absence should prompt considera-tion of another primary or secondary headache syndrome. Pathophysiology of CPH Although the pathophysiology of CPH is unknown, avail-able evidence suggests many similarities to cluster headache. Alterations in the cyclic release of catecholamine and beta-endorphins have been observed in CPH, similar to those reported in cluster headache (108). Although studies of auto-nomic function have been normal, pupillometric studies in CPH have demonstrated miosis ipsilateral to the pain, sug-gesting a partial Horner's syndrome (109). Increased levels of CGRP and VIP have been found in the cranial venous blood of CPH patients; these levels normalize after treatment (110). Segmental narrowing of the ophthalmic veins on or-bital phlebography has been reported, similar to changes seen with cluster (93). It seems that the final pathway in-volved in CPH is similar to those in cluster. The shorter attack duration, the greater attack frequency, and the selec-tive response to indomethacin may point to differences in the generation of the disorders, perhaps due to differential hypothalamic dysfunction. Treatment of CPH Indomethacin is the preferred treatment for CPH. The standard dose is 25 mg three times a day (t.i.d.), increasing to 50 mg t.i.d. after 1 week if there is no response. The response is typically prompt, occurring within one or two days. Other NSAIDs, such as naproxen, have demonstrated some effect on CPH, although less than indomethacin. Indomethacin is a nonsteroidal indoleacetic acid drug which exerts a selective treatment effect in several headache syndromes, including CPH (2,93). Indomethacin is an inhib-itor of cyclooxygenase 1 and 2, and decreases cerebral blood flow and CSF pressure. It had been speculated that the selec-tive effect of indomethacin on certain headaches is related to an effect on nitric oxide synthase (111). Indeed, nitric oxide synthase inhibitors may be effective treatments for CPH (93,111). Although there are clear similarities between cluster and CPH, acute treatments that work quite well for cluster (such as oxygen and sumatriptan) have not been con-sistently effective for CPH attacks. Hannerz and Jogestrand (112) reported a patient with bilateral CPH who responded to sumatriptan, but others have not found similar results (93). The lack of response may not necessarily imply lack of ef-fect, but rather that the time of onset for such treatments are too long to be of benefit for short-duration headaches. In contrast, verapamil seems to be effective prophylaxis for cluster and CPH. Short-Lasting Unilateral Neuralgiform Headache with Conjunctival Injection and Tearing (SUNCT) Short-lasting unilateral neuralgiform headache with con-junctival injection and tearing (SUNCT) was originally de-scribed by Sjaastad in 1989. This rare disorder predomi-nately affects men, with a gender ratio of 17 2. Patients experience short-lasting attacks of unilateral pain, most often associated with conjunctival injection and tearing. The at-tacks are much briefer than in other TACs, with a duration of 5-240 seconds. Paroxysms begin and end abruptly, reaching peak intensity within 2-3 seconds. Attacks may be numer-ous, with a mean of 28 per day in one study (11,93). In some patients, a dull discomfort persists between attacks. The pain is generally confined to the ocular/periocular region, moder-ate to severe in intensity, and often described by the patient as burning or electrical. Conjunctival injection and tearing are the most frequent autonomic features. Rhinorrhea, nasal congestion, and forehead sweating occur less commonly. An irregular temporal pattern is the rule, with symptomatic periods alternating with remissions in an unpredictable fash- HEADACHE AND FACIAL PAIN 1297 ion (114). The tearing and conjunctival injection begins 1-2 seconds after onset of pain. Some patients can precipitate attacks by touching certain trigger zones within the territory of V1-V3. Unlike trigeminal neuralgia, the attacks have no refractory period. Given the rarity of the disorder, secondary causes of SUNCT should be excluded. The most common lesion mim-icking SUNCT is a posterior fossa mass such as a vascular malformation (93,115). Ocular etiologies, such as intermit-tent angle closure glaucoma, should be considered and ex-cluded with careful slit-lamp and gonioscopic examination. Interestingly, intraocular pressures and corneal temperatures are mildly elevated during the attack, possibly reflecting parasympathetic activation with vasodilation (116). These findings, and others (2), suggest a central pathogenesis for SUNCT as a manifestation of the trigeminal vascular reflex. It is occasionally difficult to differentiate SUNCT from cluster and CPH. In contrast to cluster headache and CPH, the attacks in SUNCT are shorter and more frequent, and the repertoire of autonomic dysfunction is more limited. Women are more frequently affected in CPH. Both CPH and cluster demonstrate a consistent response to treatment, a feature lacking in SUNCT. Pathophysiology The pathophysiology of SUNCT is unknown. As with other TACs, orbital phlebography has been shown to be ab-normal in SUNCT, with narrowing of the superior ophthal-mic vein ipsilateral to the symptoms (93). Pupillary studies have revealed no abnormalities. Intraocular pressure and cor-neal temperatures are moderately elevated during attacks, likely reflecting parasympathetic activation and vasodilation (117). There have been few functional neuroimaging studies performed in patients with SUNCT, perhaps due to the rela-tive rarity of the disorder. Functional MRI revealed activa-tion in the region of the ipsilateral posterior hypothalamus in one patient (118). A central triggering mechanism with secondary activation of the trigeminal vascular and cranial parasympathetic pathways seems likely. Treatment Unfortunately, no treatments have been shown to be con-sistently and reproducibly effective for SUNCT. There have been single case reports and small case series describing a response to several agents, including gabapentin, lamotrig-ine, and sumatriptan (117). Three patients with SUNCT were successfully treated by trigeminal rhizolysis (119). How-ever, two patients were reported to fail several neurosurgical treatments for SUNCT, with residual neurologic deficits (120). Caution is always advised when assessing any therapy for a disorder characterized by an irregular temporal pattern and unpredictable remissions. Hemicrania Continua Hemicrania continua (HC) is an uncommon cause of chronic daily headache (121). This disorder is characterized by a continuous, unilateral headache that varies in intensity, without disappearing completely (122). The associated fea-tures include autonomic symptoms, ‘‘jabs and jolts'' (sharp pains lasting less than 1 minute), and migrainous symptoms (nausea, vomiting, photophobia, or phonophobia). As with CPH, HC is exquisitely indomethacin-sensitive, producing a prompt and enduring response in most patients. Some in-vestigators recommend that all patients with chronic, unilat-eral headache not attributable to other disorders receive an indomethacin trial (121). OTHER SHORT-LASTING PRIMARY HEADACHES Primary Stabbing Headache Primary stabbing headache is defined by the IHSas pain occurring as a single stab or a series of stabs confined to the head and exclusively or predominately felt in the distribution of the trigeminal nerve (orbit, temple, parietal area). The stabs may last for up to a few seconds and recur with irregu-lar frequency from one to many per day. The stabs may remain localized, or may migrate. The syndrome has carried other names over time, including ice-pick headache, idio-pathic jabs and jolts, and ophthalmodynia periodica (123). The duration of the pain ranges from 1-5 seconds, more frequently between 1-2 seconds. The attacks are usually unprovoked. There is tremendous variability in the temporal pattern of the attacks. Some patients may experience a single jab, while others may experience volleys of jabs as often as 50 times per day (2,123). Autonomic symptoms are conspic-uously absent. Visual symptoms are generally absent, al-though there is one report of transient monocular visual loss associated with primary stabbing headache (124). Primary stabbing headache is often associated with other primary headache syndromes, such as migraine. According to some studies, idiopathic stabbing pain may occur in 30-40% of migraineurs (123,125). Peres (126) identified stabbing head-ache in 41% of patients with hemicrania continua. The ab-sence of autonomic features, as well as the short duration, distinguishes this headache from cluster and CPH. Trigemi-nal neuralgia may present in a similar fashion, but uncom-monly involves V1. Trigeminal neuralgia attacks are usually provoked, and display a consistent response to carbamaze-pine, features lacking in primary stabbing headache. The pathogenesis of primary stabbing headache is un-known. Activation of the trigeminal vascular system is the most likely final pathway, but the initial events leading to activation of nociceptive centers is undetermined. It is possi-ble that the stabbing pains may result from spontaneous dis-charges of trigeminal afferents. The majority of patients with primary stabbing headache respond quite well to indomethacin, at doses similar to those used for CPH. One report describes resolution of symptoms with melatonin, perhaps implicating the hypothalamic ‘‘pacemaker'' system in pathogenesis (127). This might sug-gest similar underlying pathophysiology between primary stabbing headache and cluster headache. The lack of auto-nomic features in stabbing headache might indicate preferen-tial activation of the trigeminal vascular system with sparing of the cranial parasympathetics. Table 26.10 summarizes the clinical characteristics of short-lasting headaches. 1298 CLINICAL NEURO-OPHTHALMOLOGY Table 26.10 Clinical Features of Short-Lasting Headaches Chronic Episodic Idiopathic Cluster Paroxysmal Paroxysmal Stabbing Trigeminal Hemicrania Characteristic Headache Headache Headache SUNCT Headache Neuralgia Continua Gender (M:F) Pain Type Severity Location Attack duration Attack frequency Autonomic features Indomethcin-sensitive 9 : 1 Boring Severe Orbital/temporal 15-180 min 1-8/day Prominent 1 : 3 Throbbing/boring Severe Orbital/temporal 2-45 min 1-40/day Prominent Yes 1 : 1 Throbbing Severe Orbital/temporal 1-30 min 3-30/day Prominent Yes 8 : 1 Stabbing Moderate Orbital/temporal 5-250 sec 1/day to 30/hr Prominent No F M Stabbing Severe Any 1 sec Few to many per day Absent Yes F M Stabbing Severe V2/V3 V1 1 sec Few to many per day Absent No 1 : 1.8 Steady Moderate Unilateral Continuous May have exacerbations Occasional Yes SECONDARY HEADACHE SYNDROMES As mentioned earlier, secondary headache syndromes re-flect a wide range of intra- and extracranial pathology. In some situations, the diagnosis of secondary headache is straightforward: if a patient develops a new headache at the same time as a brain tumor, it is reasonable to conclude that the headache is secondary to the tumor. However, it is not uncommon for a patient with a primary headache syndrome (such as migraine) to develop worsening of the primary headache in close temporal relationship to a secondary cause of headache (e.g., head trauma). A secondary headache should fulfill the following criteria: 1. Another disorder known to be able to cause headaches has been demonstrated. 2. The headache occurs in close temporal relation to the other disorder and/or there is evidence of a causal rela-tionship. 3. The headache is greatly reduced or disappears within 3 months after successful treatment of spontaneous remis-sion of the causative disorder. HEAD TRAUMA AND HEADACHE Headache may occur after injury to the head, neck, or brain. The headache may be related to a structural lesion (e.g., headache attributable to intracranial hematoma), in which case management is directed toward the underlying etiology. A variety of pain patterns that resemble primary headache syndromes may develop after head injury (2). Ten-sion- type headache is the most common pattern, but mi-graine without aura and a cluster-like pattern have also been described. The role of litigation in the development of post-traumatic headache syndromes has been studied, but a con-clusive relationship has not been established. A causal rela-tionship between head and/or neck trauma is difficult to es-tablish in patients with very mild head trauma. In such cases, there is often a complex interplay between organic and psy-chosocial factors. Post-Traumatic Headache The IHSrecognizes an acute and chronic post-traumatic headache, with further division into headache associated with mild and moderate or severe head injury. There is no definite relationship between severity of headache and sever-ity of injury, although evidence suggests that post-traumatic headache is less frequent when head injury is more severe (128). The incidence of post-traumatic headache varies widely among studies, ranging from 30-90% of symptom-atic patients after head injury (2,129). The headache may be localized or diffuse, episodic or daily. Indeed, the IHScrite-ria for post-traumatic headache are not primarily concerned with the features of the headache, but with the characteristics of the head injury, and the temporal relationship of the head-ache to the trauma. By IHScriteria, the headache must oc-cur less than 7 days after head trauma or after regaining consciousness or memory (11). The headache may be ac-companied by a multitude of neurological and psychiatric symptoms, most commonly dizziness, irritability, lack of concentration, and intolerance to alcohol. This constellation of features has been called the ‘‘post-traumatic syndrome'' (2). Less common symptoms include vertigo, tinnitus, in-somnia, apathy, easy fatigability, and decreased libido. Non-specific staring spells, non-vestibular dizziness, and episodic loss of consciousness are rare sequelae of head trauma. The pathophysiology of post-traumatic headache remains much debated, with advocates for a purely neurogenic mech-anism as well as those suggesting a largely psychological etiology (130,131). It is unlikely that a single mechanism or unifying hypothesis can explain all post-traumatic head-aches. Late-onset chronic headaches following minor non-concussive head injury are difficult to explain on a purely neurogenic basis. Conversely, neurobiological factors are likely to play a role in persistent headache beginning hours to days after moderate to severe head trauma. It is probable that a combination of organic and psychosocial factors ac-count for the chronification of post-traumatic symptoms. Acute pain following head trauma may result from tissue injury to bones, blood vessels, or nerves. Keidel and Rama- HEADACHE AND FACIAL PAIN 1299 dan (132) have identified several factors that may contribute to posttraumatic headache: 1. Referred pain from nociceptive input caused by lesions of musculoskeletal, ligamentous, and other soft tissue structures, 2. Activation of meningeal nociceptive afferents due to trau-matic epidural, subdural, and subarachnoid bleeding, 3. Stretching of pain-sensitive intracranial structures from increased intracranial pressure, 4. Intracranial hypotension, and 5. Activation of the trigeminovascular system by post-traumatic venous sinus thrombosis. Diffuse axonal injury may result from angular accelera-tion of the head, and can occur even in cases of mild head injury. Several researchers have noted that mild head injury patients have more pathologic, radiologic, and electrophysi-ologic abnormalities than normal controls (130,133). How-ever, many patients with persistent symptoms have no objec-tive findings either on examination, or neuroimaging studies such as CT, MRI, SPECT, or PET. Multiple studies suggest that compensation issues and ma-lingering are not major factors in post-traumatic headaches (134). However, pre-traumatic psychopathology is more common in chronic post-traumatic headache patients than in patients with chronic TTH and low back pain. A meta-analysis of 2,353 subjects with closed head injury found that more abnormalities and disability were seen in patients with financial incentives (135). Headache following head injury should always prompt consideration of other secondary trauma-related etiologies, including intracranial hypertension and hypotension, venous sinus thrombosis, carotid or vertebral dissection, intracranial hematoma, and carotid-cavernous fistula. Many of these dis-orders will be apparent on clinical examination; others may require neurodiagnostic studies, including neuroimaging. Focal deficits on examination, new-onset syncope, impaired arousal and alertness, positional headache, and headache made worse by Valsalva maneuver all require further investi-gation. As mentioned previously, head injury may trigger or exacerbate a preexisting primary headache syndrome. The mechanism by which this occurs is unknown, although it has been suggested that sensitization may play a role: head trauma may trigger the first attack, initiating the process in predisposed individuals. Indeed, post-traumatic headache is more common in patients with a history of headache (136). Management of post-traumatic headache is dictated by the clinical syndrome, and by excluding other secondary causes. If a recognizable primary headache syndrome (such as mi-graine or cluster) has been triggered by head injury, it is treated in the usual manner for that disorder. Coexisting anx-iety and depression should be identified and treated. Tri-cyclic antidepressants are the medications most often used for post-traumatic headache, but few studies have evaluated specific drug regimens. Associated symptoms of irritability, dizziness, depression, and insomnia often respond to tricyc-lics (2). The SSRIs have been used, with some reports of effectiveness (134). Medication overuse may be a complicat-ing variable that should be addressed if treatment will be successful. Other treatment modalities, including physical therapy, exercise, good sleep hygiene, and psychotherapy, may be of value in certain patients. HEADACHE ATTRIBUTABLE TO CRANIAL AND CERVICAL VASCULAR DISORDERS Headache may be associated with vascular disorders. In some of these conditions, such as ischemic or hemorrhagic stroke, the headache is overshadowed by focal neurologic deficits. In others, such as subarachnoid hemorrhage, head-ache may be the presenting symptom. Headache may be an initial warning symptom in a number of diseases, including carotid artery dissection, giant cell arteritis, and cerebral ve-nous thrombosis. The clinician should therefore be aware of the association of headache with these potentially neurologi-cally and visually devastating diseases. The sudden onset of a new headache should always raise suspicion of a vascular etiology, particularly when the headache is severe (‘‘first and worst headache''). The relationship between headache and cerebral vascular disease was first described by Willis in 1664. Fisher reported the first extensive study of headache and vascular disease in 1968 (137). This study included clinical, arteriographic, and pathologic information as well as detailed descriptions of the headache characteristics. A number of authors have since provided detailed, and often conflicting, information regarding headache and cerebral vascular disease. Headache and Ischemic Stroke Headache accompanies stroke in 17-34% of cases (138). It is more frequently associated with strokes in the basilar rather than the carotid territory. The headache onset is usu-ally sudden, moderate in intensity, and has no defining char-acteristics. It is usually unilateral, and may last more than 24 hours. The presence or absence of head pain is of little practical use regarding stroke etiology, with two exceptions: carotid dissections frequently cause headache, while head-ache is rare in lacunar infarcts (138). The meninges investing the occipital lobe are innervated by the ophthalmic division of the trigeminal nerve; therefore, posterior circulation strokes may cause ipsilateral eye pain. Since it is not uncom-mon for patients with homonymous hemianopic defects to localize visual loss to the eye with the temporal field cut, the combination of perceived visual loss and eye pain may lead the physician to assume an ocular cause for the patient's complaints (139). Careful confrontation visual field testing or automated perimetry will clarify the diagnosis. The mechanism of head pain in stroke is poorly under-stood. As Fisher has argued (137), the basic atherosclerotic process is nearly always painless. A severe ischemic or hem-orrhagic stroke may cause pain by mass effect and stretch of pain-sensitive intracranial structures, but many headache-related strokes are relatively small. The cerebral vessels themselves are pain sensitive, and the pain is likely to be related to activation of perivascular trigeminal afferents. Va-sodilation was initially thought to be a potential mechanism for pain (140), but later studies (141,142) suggest that the vasodilation is the result of afferent activation, and the re-lease of vasoactive neuropeptides, are a secondary result, rather than primary cause, of pain. 1300 CLINICAL NEURO-OPHTHALMOLOGY Transient ischemic attacks (TIAs) may occasionally be associated with headache. Retinal TIAs are less likely to be associated with headache than hemispheric events (143). Since migraine and stroke share similar clinical profiles, it may be difficult to differentiate between migraine with aura and TIA in certain patients (144, 145). The presence of posi-tive visual phenomena such as scintillations suggests mi-graine as the most likely diagnosis, particularly given a his-tory of episodic headaches with migrainous features. Migraine aura typically lasts 20-30 minutes, in contrast to thrombo-embolic events, which are shorter (1-10 minutes). Migraine visual equivalents have a characteristic build-up or evolution, a feature lacking in TIA. Transient ischemic attacks may rarely cause positive visual phenomena (62, 146), but the duration and lack of build-up help distinguish these episodes from migraine. Headache and Subarachnoid Hemorrhage Subarachnoid hemorrhage (SAH) is by far the most com-mon cause of intense and incapacitating headache of abrupt onset, and is associated with high morbidity and mortality (2,147,148). Approximately 15-25% of patients die prior to reaching the hospital, and many survivors are left with vary-ing degrees of neurologic disability. Headache occurs in 85-95% of patients (147). The headache is sudden and ex-plosive in onset, often associated with nausea or vomiting, and is classically described by patients as the ‘‘worst head-ache of my life.'' The headache is usually followed by pain radiating into the occipital or cervical region. If the headache resolves the patient may not seek medical attention, or may do so and be misdiagnosed as having a migraine or sinus headache. This is referred to as a sentinel headache, or ‘‘warning leak,'' and may occur in 20-60% of patients (148). The headache usually occurs 2-20 days prior to the onset of major subarachnoid hemorrhage and lasts 1-2 days. In two-thirds of patients it is associated with nausea, vomit-ing, neck pain, or lethargy. In such patients, a high index of suspicion mandates further investigations, including neu-roimaging and lumbar puncture. Other presenting symptoms include seizure, lethargy, photophobia, and altered mental status. Terson's syndrome is classically characterized by vit-reous hemorrhage associated with subarachnoid hemorrhage secondary to aneurysmal rupture, and may occur in 20-40% of patients with SAH (149). Terson's syndrome has also been reported to occur with epidural and subdural hemato-mas. The presence of an otherwise unexplained vitreous or preretinal hemorrhage in a patient with new onset headache should raise concern about SAH. Headache has been reported in approximately 18% of pa-tients with unruptured aneurysms (150). The headache is generally nonspecific, and the relationship with the aneu-rysm remains debatable. A notable exception is the acute onset of a painful, pupil-involved third nerve palsy, in which case an expanding aneurysm is the default diagnosis until proven otherwise (149). Headache and Subdural Hematomas Headaches frequently occur in patients with post-trau-matic subdural hematomas (152). In a series reported by McKissock (153), headaches were reported by 11% of pa-tients with acute subdural hematomas, 53% with subacute hematomas, and 81% with chronic subdural hematomas. Be-cause many of the patients with acute subdural hematomas have alterations of consciousness, headaches in this group of patients may be underreported. Headaches in patients with subdural hematomas vary widely in severity and may be paroxysmal or constant (152). They are usually bilateral; however, when they are unilateral, they are usually on the side of the hematoma. In most patients, the subdural hematoma occurs after ob-vious closed head trauma, such as that which occurs in a fall or a severe motor vehicle accident. In other cases, however, exposure to acceleration-deceleration forces, such as those encountered while riding on a roller coaster, can tear bridg-ing cerebral veins, leading to a subdural hematoma (154). The same mechanism may be responsible for subdural hema-tomas that occur in patients with severe whiplash injuries unassociated with direct head trauma. Acute subdural hematomas are readily identified by CT scanning. Small, chronic subdural hematomas are occasion-ally missed on CT; MRI may be more sensitive in such situations. Headache and Epidural Hamartoma The classic profile of a patient with an acute epidural hematoma is one who sustains even a minor head injury with or without initial loss of consciousness followed by a lucid interval, but who then deteriorates into coma, usually within 12 hours of the injury (155). Chronic epidural hematomas, however, may produce a persistent headache that is often associated with nausea, vomiting, and memory impairment, suggesting a post-traumatic syndrome. Eventually, however, focal deficits develop, and the correct diagnosis is made (156). As with acute subdural hematomas, CT scan is the initial diagnostic test of choice. Headache and Intraparenchymal Hemorrhage Patients with acute intraparenchymal hemorrhage typi-cally experience a sudden severe headache. The hemorrhage may result from a variety of causes, including hypertension, a vascular malformation, a ruptured aneurysm, an infarction, or necrosis within a tumor. In most cases, the headache is generalized, severe, and associated with neurologic def-icits that reflect the location and size of the hemorrhage. In some cases, the clinical picture suggests the etiology of the hemorrhage. For instance, the syndrome of occipi-tal apoplexy-sudden headache, stiff neck, and a homony-mous visual field defect-is almost pathognomonic for a ruptured occipital lobe arteriovenous malformation (AVM) (157,158). Headache and Unruptured Vascular Malformati