Clinical manifestations

Presentation and course

In contrast to the headache of subarachnoid hemorrhage, which is usually explosive and of abrupt onset, the quality of pain in ischemic cerebrovascular disease varies widely. Fisher reported that most patients who had headache related to internal carotid artery disease had pain that was so minor or equivocal that it scarcely warranted mention (41). In contrast, a large proportion of patients with lateral medullary infarction reported severe, nonthrobbing headache, which in some patients was aggravated by coughing and head-shaking. Patients with vertebrobasilar ischemia usually reported throbbing or "bursting" frequently occipital headaches accentuated by stooping and straining (123).

The headache of ischemic cerebrovascular disease is more likely to be gradual in onset than sudden (28). It is usually unilateral, focal, and of mild to moderate severity (46; 117; 07); however, a significant proportion of patients (25% to 46%) may have incapacitating pain (07). Some patients present with thunderclap headache (74). The majority of patients stated that the headache had a nonspecific character and described it as either throbbing (17% to 54%) or continuous and nonthrobbing (14% to 94%) (92; 117; 07). Rarely, it may be felt as stabbing, pulsating, or having clinical features similar to intracranial hypertension (07). Headache is frequently associated with nausea (44%), vomiting (23%), and photophobia and phonophobia (25%). One study investigated headache as a symptom at stroke onset in 4431 patients with ischemic stroke <55 years (61). Headache at stroke onset was more common during ischemic stroke in females, younger patients, with greater size of the acute lesion, and if the posterior cerebral artery or vertebrobasilar system were affected. This was confirmed in another study of 129 patients (55).

The headache is usually worsened by bending, straining, or jarring the head. Transient worsening can also occur with the use of sublingual nitroglycerin. Digital compression of the superficial temporal artery on the side of the headache temporarily eases the discomfort.

The headache often accompanies the ischemic event, and this juxtaposition strongly suggests a relationship between the two events. However, in addition to the so-called onset headache, defined as temporally surrounding the neurologic deficit, several authors have suggested that the headache can either precede or follow the ischemic event by days to years (sentinel and late-onset headache, respectively) and still be related to the pathogenesis of cerebral ischemia (87).

Fisher indicated that the majority of patients have headache onset that coincides with the symptoms of ischemia in either the anterior or the posterior circulation (41). Subsequent reports estimate the frequency of onset headache to be between 8% and 34% (mean 26%) (78; 39; 40; 79; 38), but studies indicate differences of timing between headache onset and onset of neurologic deficit. Headache is rarely the initial symptom of transient carotid ischemia but commonly begins either in association with other neurologic symptoms or after the attack resolves (47). It is a constant feature in only one third of patients who experience more than one carotid transient ischemic attack; it is present in the majority of patients with recurrent vertebrobasilar ischemia (47). Patients with larger artery thrombosis develop onset headache more frequently than patients with embolic or lacunar infarcts (82). The duration of the headache is longest in cardioembolic and thrombotic infarcts, shortest in transient ischemic attacks, and of medium duration in lacunar infarction (07; 98).

Sentinel headache, which usually occurs prior to subarachnoid hemorrhage, is also not an uncommon symptom of cerebral ischemia. The reported interval between the headache and the ischemic event varies widely, ranging from a few hours to years; however, when the latter is the case, the headache is conceivably unrelated to the ischemic event. This headache is usually unilateral and focal and lasts more than 24 hours (46). The side and duration do not differ significantly among stroke subtypes. Sentinel headache occurs in 10% to 43% of patients with ischemic stroke (46). It is more frequent in cardioembolic infarcts (22%) than in transient ischemic attacks, lacunar infarcts, or thrombotic infarcts. In embolic infarctions, in particular, sentinel headache is mostly unilateral, of sudden onset, and can precede the onset of the neurologic deficit by 24 to 72 hours (42).

Lebedeva and colleagues compared 550 first-ever acute ischemic stroke patients with 192 patients who were admitted to the emergency room without acute neurologic deficits or serious neurologic or somatic disorders (67). Among the 550 patients with stroke, 94 patients (17.1%) and 12 controls (6.2%) had headache during the 7 days before stroke (p < 0.001; OR 3.9; 95% CI 1.7-5.8). Eighty-one patients (14.7%) and one control had sentinel headache within the last week before stroke. Attacks of arrhythmia during the 7 days before stroke were significantly associated with sentinel headache (p = 0.04, OR 2.3; 95% CI 1.1-4.8). This study only included Caucasians and might not be applicable to other populations (88).

The characteristics of headache and the predictors of headache at the chronic stage 12 months after ischemic stroke were investigated in a prospective observational cohort study including 102 acute ischemic stroke patients (18). Forty-five out of 89 patients with completed follow-up (51%) reported headache at the chronic stage. In most of the patients, headache was sporadic, mild, and pressure-like, with a duration of minutes to hours; 51% (n = 23/45) had characteristics of tension-type headache. Headache was a reactivation of pre-stroke headache in 33% (n = 15/45), different from pre-stroke headache in 44% (n = 20/45), and of new onset in 22% (n = 10/45).

Late-onset vascular headaches associated with ischemic cerebrovascular disease are ill-defined. According to Medina and colleagues, these are observed frequently in patients with transient ischemic attacks (52%); they begin a few days to a year after the ischemic event, and a significant proportion of them (39%) are throbbing (78). Some of these headaches may be given undeserved significance, as their association with ischemia is unclear. It is possible that the transient neurologic symptoms are due to migraine aura rather than to transient ischemic attacks, or even that ischemia activates a preexisting migraine tendency.

Headache is more frequent when ischemic events involve the posterior circulation (29% to 75%) than when they involve the anterior circulation, where headache occurs in 14% to 59% of patients (47; 78; 60; 80; 117; 57; 63; 79; 51). Headache with minor ischemic stroke was positively associated with cortical infarcts (OR 1.78, 95% CI 1.31 - 2.41) (51). Headache occurs with an intermediate frequency when the vascular topography is indeterminate (33%). Headache is also more common when there is cortical involvement (56%) than when there are subcortical infarctions (26%) (07). Arboix and colleagues recorded headache in 484 patients with lacunar strokes; 9.3% presented headache within a 72-hour interval of stroke onset. Intensity of headache was mild in severity and poorly localized (05).

Jorgensen reported that headache was lateralized in 46% of patients, being ipsilateral to the stroke in 68% of these patients and contralateral in 32% (57).

Headache occurs in 10% to 31% of patients with symptomatic atherothrombotic disease of the internal carotid artery (105). However, most patients experience no pain regardless of the severity and duration of the neurologic symptoms. The headache is usually frontal and lateralized, but frontal nonlateralized, fronto-occipital, and cervico-occipital pain has also been reported. Discomfort over the carotid area in the neck may occur but is rare (41).

The typical pain of middle cerebral artery stem thrombosis is located behind, in, and above the corresponding eye. It is usually steady and may precede, accompany, or develop during ischemia (41). In contrast, middle cerebral artery embolism is more likely to produce headache on the side of the head above the temple (41). Headache occurs in 10% to 39% of patients with middle cerebral artery occlusive disease (105) and is more common with thrombosis than with embolism (41).

In anterior cerebral artery occlusive disease, the incidence of headache ranges from 0% to 18% (15). The headache is usually severe and bifrontal, unilateral frontal, or even bioccipital.

Between 64% and 90% of patients with posterior cerebral artery distribution ischemia develop headache (41; 17). It is more frequent with thrombosis than with embolism (41). The headache is usually frontal and lateralized, but less often is frontal nonlateralized, fronto-occipital lateralized or nonlateralized, occipital lateralized or nonlateralized, over the vertex, or in a hat-band distribution (41). Rarely, it may be throbbing and worsen with coughing or head shaking. Altiparmak and colleagues investigated 200 patients had cerebellar ischemia and 22 patients (11%) had headache at initial presentation (04). Patients presenting with a headache predominantly had left hemispheric (vermian > hemispheric > peduncular) involvement.

Reversible cerebral vasoconstriction syndrome is characterized by the association of severe headache with neurologic symptoms (56). Focal neurologic deficits can last for hours, even days, and are fully reversible (35; 45). Angiography shows “string and beads” sign of cerebral arteries. Reversible cerebral vasoconstriction syndrome is treated with nimodipine. This syndrome can also appear 1 to 2 weeks after delivery, and lead to severe headache and stroke (43).

The majority of patients (68%) with vertebral artery occlusion and lateral medullary syndrome report head discomfort (41). Two different types of pain are described. The first and less frequent, considered typical of the lateral medullary syndrome, is located in the eye, nose, and cheek and is probably related to ischemia of the nucleus of the descending root of the trigeminal nerve because it is usually succeeded by numbness. The second is probably of arterial origin and is usually suboccipital or occipital and lateralized in location; however, occipital nonlateralized, occipitofrontal, dull frontal, nuchal, or generalized locations have been recorded (89). Coughing and head shaking can worsen the pain. Neck ache, centered around the C5 level, is probably secondary to occlusion of the cervical portion of the vertebral artery. Brainstem ischemia can result in indomethacin responsive hemicrania continua (114).

Headache is a frequent accompaniment of symptomatic basilar artery occlusive disease. It occurs in 21% to 53% of patients with major or minor basilar syndromes (123). Most often, it is occipital and can be either lateralized or nonlateralized. Less frequently, it is localized to occipitofrontal or frontal distributions, lateralized or nonlateralized, and rarely, it is described as band-like. Occasionally, there is associated occipital tenderness and neck stiffness. The pain is throbbing or banging, and it is aggravated by postural changes, stooping, and straining. Commonly, headache is related to the neurologic deficit, but it may persist longer than the other symptoms or even occur independently.

Between 50% and 66% of patients with subclavian steal syndrome complain of pain located either in the cervico-occipital or mastoid regions. It is usually throbbing and precipitated by strenuous exercise and not associated with neurologic deficit (53).

Prognosis and complications

The prognosis depends on the primary cerebrovascular condition that caused the headache. Jorgensen and colleagues found that headache has no independent relationship to neurologic outcome and does not alter the overall outcome of ischemic stroke (57). In contrast, the Taiwan Stroke Registry reported that 7.4% of 11,523 patients with first ever stroke reported new headache (23). Compared with patients without onset headache, those with onset headache had a lower frequency of stroke in evolution (4.5% vs. 6.7%; adjusted relative risk, 0.64; 95% confidence interval, 0.52-0.79), greater improvement in National Institutes of Health Stroke Scale score on discharge (0.08 vs. -0.20; P=0.02), higher mean Barthel index scores (86.5+/-20.0 vs. 83.9+/-23.3; adjusted difference, 1.43; 95% confidence interval, 0.28-2.89), and a lower frequency of modified Rankin scale higher than 2 (27.6% vs. 31.5%; adjusted relative risk, 0.85; 95% confidence interval, 0.72-0.95) at 1-month follow-up. There was also a trend for better functional outcome in 3- and 6-month follow-up. In a study of 1411 acute stroke patients, the Dijon Stroke Registry found increased early mortality associated with headache at stroke onset only in patients with cerebral hemorrhage, but not in patients with ischemic stroke (01).

There are no clinical features in patients admitted to an emergency room with headache, which would predict an increased risk of ischemic stroke (70).

Clinical vignette

A 35-year-old woman was admitted to the emergency room with complaints of unilateral throbbing headache accompanied by nausea, photophobia, and phonophobia. She also reported visual disturbances in the left visual hemifield and left-sided disturbances of sensation. Her history revealed that she had suffered from migraine with aura since the age of 18 years, with three to five attacks per year. The attending emergency physician suspected a migraine aura and treated the headache with aspirin. The neurologist recorded the history again and learned that in contrast to the past, during this event the neurologic deficits did not develop slowly, within 20 minutes, but were fully developed within a minute. The MRI including diffusion-weighted imaging showed a right posterior circulation stroke. The time interval between onset of symptoms and the correct diagnosis was longer than 3 hours. Therefore, systemic thrombolysis could not be performed. The history further revealed the existence of several vascular risk factors such as smoking, oral contraceptives, obesity, and lack of exercise.

This case shows that the development of aura symptoms usually is slow and not abrupt as in a case of posterior circulation infarct. The headache might be that same as that seen in a typical migraine attack and does not help to differentiate migraine aura, ischemic stroke, or even cerebral hemorrhage.

Biological basis

Etiology and pathogenesis

The headache of ischemic cerebrovascular disease is classified as a secondary headache because there is a clear underlying etiology: ischemic stroke or transient ischemic attack. Therefore, only its frequency will be analyzed with respect to the etiology of the underlying cerebral ischemia. In many patients, headache in the acute phase of stroke seems to be a reactivation of preexisting primary headache (116). Seifert and colleagues assessed patients with acute ischemic stroke (n = 100) by brain MRI at 3 T including diffusion weighted imaging (102). Fifty patients with stroke and headache as well as 50 patients with stroke but no headache symptoms were included. Infarcts were manually outlined, and images were transformed into standard stereotaxic space using nonlinear warping. Voxel-wise overlap and subtraction analyses of lesions as well as nonparametric statistics were conducted. Between the headache group as well as the nonheadache group there was no difference in infarct volumes, in the distribution of affected vascular beds, or in the clinical severity of strokes. The headache phenotype was tension-type-like in most cases. Subtraction analysis revealed that in headache sufferers, infarctions were more often distributed in two well-known areas of the central pain matrix: the insula and the somatosensory cortex. The insular cortex is a well-established region in pain processing. The results suggest that, at least in a subgroup of patients, acute stroke-related headache might be centrally driven. In a further post hoc analysis of voxel-based MRI in 49 patients with poststroke headache, intensity of headache was associated with lesions of the posterior insula, operculum, and the cerebellum (103).

Atherothrombotic cerebral ischemia. Headaches are frequently associated with atherothrombotic cerebrovascular disease. Gorelick and colleagues reported no statistically significant difference in the frequency of onset headache among patients with disease of the extracranial carotid artery, the carotid siphon, the middle cerebral artery, or the carotid siphon and middle cerebral artery in tandem (46). In the Harvard Cooperative Stroke Registry, onset headache occurred in 12% of patients with large artery thrombotic infarctions, whereas sentinel and late-onset headaches were reported in 10% and 9%, respectively (82). Fisher reported headache in 31% of patients with internal carotid artery stenosis or occlusion and in 21% of patients with middle cerebral artery thrombosis (41).

Cerebral embolism. It is a common belief that headache is more frequent in embolic infarcts or that headache at the onset of neurologic deficit in ischemic cerebrovascular disease indicates embolism. In a large series of cases of cerebral embolism, McDowell reported onset headache in 18% of patients (77). It was mild in 10% and severe in 8%. In the Harvard Cooperative Stroke Registry, onset headache occurred in 9% of the embolic infarctions, whereas sentinel and late-onset headache occurred in 5% and 11% of the patients, respectively (82). These figures are similar to those associated with atherothrombotic stroke. Fisher reported that only 14% of patients with middle cerebral artery embolism had headache, in contrast to 21% of those with middle cerebral artery thrombosis (82). Posterior cerebral artery thrombotic infarcts also were more likely to cause headache than embolic posterior cerebral artery occlusion. Thus, these data indicate that, in embolic infarcts, headache is equally or perhaps even less frequent compared with atherothrombotic. In a CT-angiography study the clot extent in intracranial vessels was not related to headache (115).

Lacunar infarcts. There are conflicting reports regarding the incidence of headache in patients with lacunar infarcts. Fisher reported headache in only 4 of 70 patients with pure motor hemiplegia and in only 1 of 36 patients with pure sensory stroke (41). Atkinson and Appenzeller suggested that small intracerebral arterioles are not innervated, and, therefore, disease of these vessels would not be expected to cause headache (11). Several studies indicated that headache is an infrequent accompaniment of lacunar infarction, occurring in 3% to 6% of patients (82). Other studies indicated that headache is not such an uncommon occurrence in lacunar infarction, because 10% to 23% of their patients reported onset headache (92; 101). Notably, 27% of the patients with headache accompanying small deep infarcts also reported symptoms suggestive of cortical dysfunction, such as aphasia or visual field deficits (60). Thus, the difference in the frequency of headache with lacunar infarction in the above-mentioned series could be related, at least in part, to different definitions of this condition. In a cohort of 387 patients with neuroimaging-proven acute lacunar infarction collected from a prospective, hospital-based stroke registry over a 12-year period, 43 patients (11.1%) presented with headache within a 72-hour interval of stroke onset (06). Headache was more common in deep brain gray matter or brainstem lacunar infarction than in supratentorial white matter lacunar infarction (14.9% vs. 8%, P< .033), but lacunar infarctions in the supratentorial white matter less frequently had absence of limitation at discharge (15.1% vs. 25.1%, P< .013). In deep brain gray matter or brainstem lacunar infarction, early neurologic recovery decreased from 26.2% to 19.2% when headache was present at stroke onset. In the multivariate analysis, dysarthria-clumsy hand and absence of headache in deep brain gray matter or brainstem lacunar infarction were independent predictors of favorable outcome.

Headache after stroke can be induced by drug treatment. Nitrous oxide donors frequently cause headache. Dipyridamole given in combination with acetylsalicylic acid can cause headache in the first days of use (33).

The mechanism of headache in ischemic cerebrovascular disease is poorly understood. The basic atherosclerotic process, regardless of its site, is always or almost always painless (41). This observation is supported by the observation that headache after stroke is less prevalent in patients with atherosclerosis than in patients without atherosclerosis on CT angiography (115). The entire parenchyma of the cerebrum and cerebellum, including the intraparenchymal vessels, is insensitive to all forms of stimulation (96). Massive hemorrhage or severe edema complicating an ischemic infarct can cause head pain by displacing and stretching pain-sensitive intracranial structures. Many ischemic infarcts are not massive, however, and are not complicated by hemorrhage or severe edema but are accompanied by headache.

The headache associated with ischemic cerebrovascular disease is presumably vascular in origin, arising from either the intracranial or the extracranial vessels. By applying direct mechanical or electrical stimulation intraoperatively, Ray and Wolff mapped the pain-sensitive intracranial structures (96). The following arteries were pain-sensitive:

The pial arteries over the superior and lateral convexities of the cerebrum and the cerebellum were insensitive to pain.

Another possible mechanism is increased blood pressure in the acute phase of stroke. Hong and colleagues found that elevated systolic blood pressure correlated with headache within the first 24 hours after stroke onset (54).

Moskowitz and colleagues pointed out that the circumscribed unilateral headache and pain referred to the cutaneous receptive field of the first trigeminal division indicated an important role for the trigeminal nerve in pain transmission and possibly blood flow control (83; 84). Moskowitz and his coworkers demonstrated that the pial nerve fibers were of trigeminal origin and that the perivascular nerve fibers contained vasoactive neuropeptides (eg, calcitonin gene-related peptide, substance P), which, on release into the vessel wall, increase blood flow and vascular permeability. The origin and distribution of the perivascular afferent fibers explain several unique features of vascular headache. For example, the predominantly ipsilateral distribution of trigeminal fibers explains the strictly ipsilateral distribution in many vascular headaches. In addition, the bilateral innervation of certain vessels (eg, anterior cerebral artery) explains the bilateral or even contralateral location of the headache in diseases affecting these vessels. Moreover, the dual innervation of the superior cerebellar artery and the rostral basilar artery (ie, from the upper cervical roots and the trigeminal fibers) provides an anatomical explanation for the coexistence of occipital and frontal headaches. Finally, the observation that some dural and pial arteries receive divergent axon collaterals from single trigeminal neurons may account for the difficulty in distinguishing the source of pain in vascular headache. The same sensory ganglia would discharge with appropriate stimulation in both circulations.

Mitsias and colleagues analyzed prospectively collected data relevant to headache occurring at ischemic stroke onset in consecutive patients included in the Henry Ford Hospital Stroke Data Bank (81). Three hundred and seventy-five patients had complete headache and clinical data sets and were included in the analysis (headache, n=118; no headache n=257). Multivariate analysis revealed that the independent predictors of headache were infarct in the distribution of the posterior circulation (P=0.0076, odds ratio 2.15), absence of history of hypertension (P=0.0106, odds ratio 0.48), and treatment with warfarin at the time of the index stroke (P=0.0135, odds ratio 4.89). The occurrence of headache at onset of ischemic stroke was determined by posterior circulation distribution of the ischemic event, absence of a history of hypertension, and treatment with warfarin at the time of the index stroke. These results suggest that preserved elasticity and maintenance of the intracranial vasculature in a relaxed state, in combination with coagulation system derangements, and activation of dense perivascular afferent nerves play a role in the pathogenesis of onset headache.

These discoveries indicate that headache in cerebrovascular disorders is mostly related to electrochemical or mechanical stimulation of the trigeminovascular afferent system. The evidence that posterior circulation ischemic events are more often associated with headache and the suggestion that the posterior circulation is more densely innervated by the trigeminovascular system support the above theory. What triggers the trigeminovascular system in ischemic stroke remains to be determined.

Other possible triggers could originate intravascularly. Circulating hormones, biogenic amines, and antiphospholipid antibodies have all been considered in the pathogenesis of headache in ischemic stroke, but their precise role has yet to be determined. Platelet aggregation, the products of the "release reaction" including serotonin and prostaglandins, and the primary and secondary effects of these products on pain-sensitive vessels also have been incriminated in migraine. A similar mechanism may operate in ischemic cerebrovascular disease, where platelets play a more significant and prominent role.

Castillo and colleagues reported that elevated plasma and CSF glutamate levels and decreased plasma and CSF taurine levels were found in patients with headache during the acute phase of ischemic cerebrovascular disease (20). These findings suggest that during cerebral ischemia the release of excitatory amino acids in combination with the decrease of taurine (which may carry postsynaptic inhibitory effects) may lead to excessive spontaneous depolarization and a state of neuronal hyperexcitability, possibly promoting pain mechanisms. The excessive release of glutamate may also modify the cerebral circulation by intermediate mechanisms, acting as a major stimulus for the production of nitric oxide and resulting in local increases in cerebral blood flow and arteriolar vasodilation.

None of the above mechanisms can explain the occurrence of headache in lacunar strokes, as the intracerebral arterioles are not innervated, and disease of such vessels would not be expected to cause headache. Nevertheless, Koudstaal and colleagues found that 27% of patients with headache accompanying small deep infarcts also reported symptoms suggestive of cortical ischemia. This suggests that the infarct was probably caused by occlusion of the stem of one major cerebral artery, resulting in ischemia of the territory of the deep perforators so that the headache occurred because the large vessel was involved (60).

De novo headache after thrombectomy. Gallo and colleagues performed a prospective evaluation of clinical features of headache after thrombectomy (44). One hundred seventeen patients were included (52.1% females). Most of the patients had an anterior circulation artery occlusion (91.5%). Thirty-one patients (26.5%; 95% CI 18.8–35.5%) had headache related to thrombectomy, and this headache was associated with a history of primary headache (p = 0.004). No differences were seen between patients with or without post-thrombectomy headache in relation to sex, initial NIHSS score, or the type or complexity of the procedure. Headache was usually moderate, ipsilateral to the occluded artery and usually lasted less than 48 hours. In conclusion, almost one-third of patients with ischemic stroke who undergo endovascular thrombectomy experience headache in the first 24 hours, occurring more frequently in patients who had a previous history of headaches.

Epidemiology

A meta-analysis calculated the pooled prevalence of ischemic stroke-related headaches in Chinese patients (Xie at al 2022). Ninety-eight studies were eligible for inclusion. The overall pooled prevalence of ischemic stroke-related headache was 18.9%. Subgroup analysis showed that the prevalence of ischemic stroke–related headaches was higher among studies using a self-reported diagnosis of headache (18.9%; 95%CI, 8.9% to 40.2%), and those focused on age greater than or equal to 55 years (19.7%; 95%CI, 14.9% to 25.9%) and rural settings (24.9%; 95%CI, 19.7% to 31.6%). The prevalence of pre-onset headache (13.9%) and tension-type headache (15.5%) and was higher compared with other types of headache. A prior history of headache (OR = 3.24; 95%CI, 2.26 to 4.65.), female gender (OR = 2.06; 95%CI, 1.44 to 2.96.), midbrain lesions (OR = 3.56; 95%CI, 1.86 to 6.83.), and posterior circulation stroke (OR = 2.13; 95%CI, 1.14 to 4.32) were major risk factors.

Transient ischemic attack and minor stroke. Headache has been reported in 6% to 65% of patients with transient ischemic attack and minor stroke (47; 78; 39; 40; 91; 51). A meta-analysis of 20 prospective studies showed a pooled prevalence rate of headache after ischemic stroke of 14% (51). In a retrospective analysis, Grindal and Toole reported headache in 25% of patients with transient ischemic attacks (47). The authors considered this figure to be an underestimate because of the retrospective nature of the study and because only 21% of the patients gave a definite negative history of headache. In contrast, Medina and colleagues reported headaches in 65% of 34 prospectively evaluated patients with transient ischemic attacks (59% in the anterior and 75% in the posterior circulation) (78). Onset headache occurred in only 44%, however. In a prospective study, Edmeads found headache in 26% of patients with carotid transient ischemic attack and in 17% of patients with vertebrobasilar transient ischemic attack (36). Portenoy and colleagues reported headache in 10 of 28 (36%) prospectively evaluated patients with transient ischemic attack (92). In the study by Loeb and colleagues, headache occurred in 30% of transient ischemic attack patients; there was no difference between headache and nonheadache patients with regard to gender and age (72). The headache prevailed in patients with vertebrobasilar transient ischemic attack. In the Dutch transient ischemic attack prospective trial, headache occurred in 18% of 3126 patients with acute cerebral and retinal ischemia and was equally common among patients with minor stroke (19%), reversible ischemic neurologic deficit (18%), and transient ischemic attacks (16%) (60). Arboix and colleagues found headache in 39% of 31 patients with transient ischemic attack (07). The mean age of this study population was 66±13.5 years.

Transient ischemic attack and completed ischemic stroke. In a large series including patients with both transient ischemic attacks and cerebral infarcts, Fisher found that 31% of patients with internal carotid stenosis or occlusion (proven by pathologic or arteriographic examination), had headache, independent of whether they had transient ischemic attacks, a minor stroke, or a major neurologic deficit (41). Headache was reported by 44% and 35% of patients with basilar territory infarcts or transient ischemic attacks, respectively. In the Harvard Cooperative Stroke Registry, headache occurred at the onset of the ictus in 9% of patients with cerebral embolism, in 12% of those with large artery thrombosis, and in 3% of patients with lacunar infarcts (82). Headache preceding the ischemic event occurred in 10% of patients with large artery thrombosis, in 5% with embolic infarcts, and in 6% with lacunar infarcts, whereas the figures for headache following the event were 9%, 11%, and 2%, respectively. Edmeads found headache in 25% of prospectively evaluated patients with either transient ischemic attacks or ischemic infarcts admitted to the stroke unit of a university medical center (36). Patients with aphasia or other factors that prevented reliable determination of whether headache occurred were excluded. Portenoy and colleagues reported headache in 29% of consecutive patients with ischemic infarcts and in 36% with transient ischemic attacks who were evaluated prospectively at two teaching hospitals in New York (92). After analysis of data derived from a stroke registry at two inner-city referral hospitals in Chicago, Gorelick and colleagues reported onset headache in 17% of patients with ischemic stroke (46). Patients with vertebrobasilar occlusive or embolic disease and nonatherosclerotic strokes were excluded. Vestergaard and colleagues reported headache in 26% of 214 patients with cerebral infarction (117), whereas Arboix and colleagues found headache in 32% of 195 patients with ischemic stroke (07). Hansen and colleagues followed 299 consecutive stroke patients for 6 months and observed new headaches in 13.1% (49). In a large prospective study including 867 patients with ischemic and hemorrhagic stroke, Jorgensen and colleagues indicated that headache occurred in 25% of the patients with ischemic stroke (57). It was more frequent when the vertebrobasilar circulation was involved (37%), whereas fewer patients (26%) with carotid territory ischemia experienced headache. Women developed headache more frequently than men (31% vs. 25%). Patients with headache and ischemic stroke tended to be younger than patients without headache (71.4 ±12 years vs. 74.1 ±11 years). In a community study, 27.4% of 402 patients with ischemic stroke complained of headache (95). In a study of 2196 patients with transient ischemic attack or ischemic stroke, headache occurred in 27% of the patients (110). In a multivariate analysis, headache at stroke onset was correlated with female gender, a history of migraine, younger age, and lower blood pressure at admission. In a registry study in Israel with 2166 patients with acute stroke, younger age, female gender, posterior circulation involvement, and prior headache history were predictors of headache after stroke (91; 02).

Hansen and colleagues followed 256 patients 3 years after a stroke for persistent novel headache following stroke (50). Twelve percent (26/222) of patients reported new persistent headache. The headache was tension-type-like in 50%, migraine-like in 31%, and fulfilled criteria for medication overuse in 6%. In the study by Lebedeva and colleagues at 3 months, 61 of 529 patients with first-ever ischemic stroke (30 women and 31 men) (11.5%) had headache 3 months after stroke: 34 had a new type of headache, 21 had a headache with altered characteristics, and six patients had a headache without any changes (68). Therefore, 55 (10.4%) patients had a persistent headache attributed to ischemic stroke. The clinical features included less severity of accompanying symptoms, slowly decreasing frequency, and development of medication overuse headache in one-third of the patients.

The difference in the frequency of headache among the quoted studies is related to several factors, such as the study population, stroke subtype, nature of the study (prospective vs. retrospective), method of data collection, and inclusion criteria. With the exception of those studies that included only patients with transient ischemic attack, the frequency of headache in patients with ischemic cerebrovascular disease appears to be underestimated. Patients with language dysfunction, altered mental status, or other factors preventing reliable determination of a headache complaint were excluded from most studies. In addition, patients with severe sensory loss, preventing or modifying the headache; memory loss; and pain asymbolia as part of the acute stroke syndrome may negate the complaint of the headache, and this results in an overall lower frequency of headache. On the other hand, pre-CT scanning studies used clinical criteria alone to differentiate ischemia from hemorrhage; therefore, some hemorrhagic strokes could have been included in the ischemic category, thus, increasing the estimated frequency of headache in that group.

An interesting aspect of migraine and stroke is the observation that in 124,558 surgical patients, those with migraine had a higher 30-day peri- and post-operative stroke risk (112).

Prevention

Prevention of ischemic cerebrovascular disease, with antiplatelet drugs and anticoagulants, depending on the mechanism and vascular structures involved, would be expected to decrease the occurrence of this type of headache. Whether pretreatment with antiplatelet agents or antimigrainous medications (such as beta-blockers, calcium antagonists, or anticonvulsants) alters the frequency of headache at the onset of stroke remains to be determined.

Differential diagnosis

Headache with neurologic deficits and cerebrospinal fluid lymphocytosis (HaNDL). Headache with neurologic deficits and cerebrospinal fluid lymphocytosis (HaNDL) can present with stroke-like manifestations before the diagnosis is verified by lumbar puncture (08). The condition has a good prognosis with full recovery from neurologic symptoms, and neuroimaging shows normal results. Perfusion CT might help to differentiate HaNDL from acute ischemic stroke (90).

Drugs. Drugs used in the acute phase of stroke and for secondary stroke prevention can cause headache. This is true for nitrates used for angina pectoris and dipyridamole used for secondary prevention (33; 71). In a randomized trial, transdermal glyceryl trinitrate resulted in headache in 15% of patients with acute stroke compared to 0% with placebo (94). Headache due to dipyridamole use is more frequent in patients suffering from migraine (62) and can be prevented in part by an initial dose reduction and titration (22; 32; 34). The female sex, transient ischemic attack, absence of hypertension, and nonsmoking are predictors of headache from aspirin plus dipyridamole (48; 73). However, a randomized open study with 114 patients failed to show that slower than standard dose escalation reduces drug-induced headache frequency (29). Interestingly, stroke patients who develop dipyridamole-induced headache have a better long-term prognosis for recurrent stroke, indicating preserved cerebrovascular function (24).

Dissection. Carotid or vertebral artery dissection is a common cause of stroke in persons younger than 50 years of age (27; 12; 19). Neck and facial pain radiating on the temporal and frontal region is typical for carotid dissection, whereas isolated neck pain may occur in vertebral artery dissection (14; 13; 104; 108; 128; 118). Headache in intracranial vertebral dissection has an acute onset and is of persistent quality. Pain has a throbbing character and severe intensity on the ipsilateral side (59). Migraine increases the risk of dissections (113; 03; 97). Headache occurs in 88% of patients and may precede neurologic symptoms in 53% of the patients (100). Headache can be the only symptom in arterial dissection (10; 76). The headache that occurs with vertebral artery dissection can mimic cluster headache (65). Headache is more frequent in vertebral artery dissection than in stroke to large vessel disease of the vertebral arteries (69). In the CADISP study of 982 consecutive patients, headache was more frequent in internal carotid artery dissection than in vertebral artery dissection (26).

Intracerebral hemorrhage. Primary intracerebral hemorrhage frequently presents with headache and focal neurologic deficit. Headache is a leading symptom in cerebral hemorrhage. The likelihood ratio in a systematic literature review that headache predicts cerebral hemorrhage was 2.9 (99). Aneurysmal subarachnoid hemorrhage commonly presents with severe headache and focal neurologic deficit. Prior to rupture, expanding aneurysms can produce localized head pain in areas commonly affected by the headache of ischemic cerebrovascular disease. Cerebral arteriovenous malformations may also present as headache associated with focal neurologic deficit. The prevalence of headache associated with arteriovenous malformations ranges from 14% to 79%, with a higher headache prevalence in occipital arteriovenous malformations (37).

Other conditions. Early in its course, cerebral neoplasia may present with localized headache and neurologic deficit. Partial seizures, particularly those associated with postictal focal neurologic deficit (Todd paralysis), must be considered in the differential diagnosis.

Rare familial disorders such as the mitochondrial encephalomyopathies (MELAS and its variants) present with focal neurologic deficits, headache, and seizures. In these instances, muscle biopsy can be of particular diagnostic importance.

Confusing conditions

The differential diagnosis includes migraine with aura (especially hemiplegic migraine) and the major differential diagnoses of headache related to transient ischemic attacks. A subgroup of migraine patients with unilateral motor symptoms and migraine will present with motor weakness outside of migraine attacks. About 40% of these patients are diagnosed with stroke (127). The mode of evolution of the neurologic deficit and the accompanying headache, the history of prior similar attacks, the possible family history of a similar problem, and frequently the negative diagnostic workup will point toward that diagnosis. Computed tomographic perfusion imaging might help to differentiate migraine aura from a transient ischemic attack (107). Migraine-induced stroke enters the differential diagnosis in cases where the neurologic deficit is long-lasting. In these cases, the patient has previously fulfilled criteria for migraine with neurologic aura, and the present attack is typical of previous attacks. However, neurologic deficits are not completely reversible within 7 days, neuroimaging demonstrates ischemic infarction in the relevant area, and other causes of infarction are ruled out by appropriate investigations (121).

Diagnostic workup

There is no specific diagnostic workup for the headache itself. The diagnostic evaluation should be directed toward identifying the location, vascular distribution, and mechanism of the ischemic cerebrovascular disease. This should include hematological testing, with particular attention to red cell count and hemoglobin as well as platelet function. If necessary, especially in young patients or patients with cryptogenic stroke, testing for anticardiolipin antibodies, lupus anticoagulant, antithrombin III, and protein C and S should be considered.

Neuroimaging studies (CT, MRI, MRA, or CTA) are necessary to evaluate the possibility of intracerebral hemorrhage, subarachnoid hemorrhage, aneurysm, or cerebral neoplasm and to evaluate the location, vascular distribution, and extent of the ischemic infarct (126).

Cerebrospinal fluid analysis may exclude subarachnoid hemorrhage, central nervous system infection, or primary CNS angiitis.

Vascular studies, including B-mode carotid and transcranial ultrasonography, magnetic resonance angiography, or even conventional cerebral angiography should be considered, according to the clinical presentation.

Cardiac evaluation with electrocardiography, Holter monitor, and transesophageal echocardiography may exclude or confirm a cardiogenic or aortic source of cerebral embolism.

Management

No specific information is available concerning the specific treatment of headache associated with ischemic cerebrovascular disease. When hemorrhage is excluded, headache can be treated with acetylsalicylic acid or acetaminophen. Triptans are contraindicated despite the fact that they do not increase the risk of stroke. Gepants and monoclonal antibodies against CGRP or the CGRP receptor should be avoided in patients with risk of recurrent transient ischemic attack or ischemic stroke (30; 75; 31). Anti-CGRP drugs might inhibit the opening of collaterals in situations of brain ischemia (85). If antihypertensive treatment is indicated in stroke patients with a history of migraine, beta-blockers are preferred over calcium-channel blockers (119).

Special considerations

Anesthesia

No specific anesthesia requirements have been reported. These depend on the primary cerebrovascular condition.