++
Migraine is one of the most common primary headache disorders
and is characterized by unilateral, throbbing headaches associated
with nausea, vomiting, photophobia, and phonophobia. Prior to the
onset of headache, some migraineurs experience transient focal neurologic
symptoms, which may include visual disturbances, unilateral numbness,
and weakness, as well as language dysfunction. Probably because
of these neurologic symptoms as well as the intensity the headache,
the research and speculation surrounding the pathophysiology of
migraine has been the most intensive of all primary headache disorders.
The speculation that has arisen around migraine has greatly influenced
the discussion of pathophysiology of other headache syndromes. Traditional
theories of migraine pathogenesis fall into two categories: vasogenic
and neurogenic.
++
In the late 1930s, Dr. Harold Wolff and coworkers observed that
(1) extracranial vessels became distended and pulsated during migraine
attacks in many patients, implying that dilation of cranial vessels
might be important in migraine; (2) stimulation of intracranial
vessels in awake patients resulted in an ipsilateral headache; and
(3) vasoconstricting substances, such as ergotamine, could abort
headaches, whereas vasodilatory substances, such as nitrates, could
trigger migraine attacks. Based on these observations, it was theorized
that intracranial vasoconstriction was responsible for the aura
of migraine and that the subsequent headache resulted from a rebound
dilation and distention of cranial vessels and activation of inflamed
perivascular sensory neurons.
++
The competing neurogenic theory held that migraine is a brain
disorder based on an altered cerebral susceptibility to migraine
attacks and that the vascular changes occurring during a migraine were
the result rather than the cause of the attack. Advocates of the
neurogenic theory pointed to the neurologic symptoms, both focal
(in the aura) and vegetative (in the prodrome), that often are prominent
components of migraine attacks and cannot be explained on the basis
of vasoconstriction within a single neurovascular territory. The
expanding nature of the visual and sensory symptoms during migraine
aura has led to speculation that the phenomenon of spreading depression might
underlie the aura.23 Spreading depression is a
wave of neuronal hyperexcitation followed by suppression that is observed
to move across areas of contiguous cortex in experimental animals
after chemical or mechanical perturbation.24 The
speculation that spreading depression might be important in migraine
aura was reawakened when Olesen, Lauritizen, and their colleagues
employed intraarterial xenon 133 (133Xe) blood flow
techniques to investigate the hemodynamic changes occurring during
aura-like symptoms induced during carotid angiography. Olesen and
coworkers reported that aura symptoms were accompanied by reductions
in cerebral blood flow, usually in posterior regions of the brain.25,26 Some
studies reported a transient increase in blood flow prior to the
blood flow reductions as well as an apparent anterior spread of
the blood flow decrements, which moved across neurovascular boundaries.26 The
estimated rate of the spread was about 2 to 3 mm per minute,26 although
the accuracy of this rate has been questioned because of the convoluted
nature of the human cerebral cortex. The estimated reductions in
blood flow observed in these 133Xe blood flow studies ranged
from 17%27 to 35%,25 well
above the threshold (i.e., > 75%)
for frank ischemia, and were therefore termed spreading oligemia.
However, some researchers have speculated that the artifact of Compton
scattering might account for both an underestimation of the magnitude
of blood flow reduction and the apparent spreading pattern of the
blood flow change.28
++
If applied rigidly, neither of these traditional theories completely
explains the clinical symptomatology of migraine. It is likely that
migraine is not a disease per se but, rather, a syndrome in which
acute attacks occur when one or more triggering environmental events
interact with a vulnerable nervous system. Why certain individuals
possess this vulnerability to migraine attacks is not fully understood
but is likely a result of a combination of genetic and acquired
factors. There is great variability among the environmental triggers
that are potentially capable of inciting a migraine attack. Most
migraineurs are aware of several to which they are sensitive. Those
triggers most commonly reported include exposure to certain foods
or food additives; certain types of physical exertion; alteration
of usual sleep patterns; increased personal or professional stress;
hormonal fluctuation; unaccustomed fasting; exposure to glaring,
flickering lights, or strong smells; and changes in weather patterns
or barometric pressure. The triggers for attacks vary widely from one
individual to another and may change with time for an individual
migraineur. The mechanism by which these provocative factors start
the attack are not well understood. Neither the biochemical nature
nor the exact site of migraine initiation is known, but recent advances
in functional neuroimaging are beginning to yield some clues.
++
A recent investigation performed during the first 6 hours of
nine spontaneous attacks of migraine without aura using PET has
led to speculation that a so-called migraine generator may exist
in the proximal brainstem.29 In this study, a significant
increase in regional cerebral blood flow (rCBF) was observed in
the anterocaudal cingulate cortex and in visual and auditory associative
cortices. In addition, an 11% increase in rCBF was seen
in medial brainstem structures over several planes, slightly contralateral
to the headache side. Activation in the medial brainstem but not
the cingulate and auditory association cortices persisted even after
effective treatment of the headache with 6 mg of subcutaneous sumatriptan.
Activation within the medial brainstem was not observed during a
subsequent study during a headache-free interval nor was it observed
in another series after subcutaneous injection of capsaicin in the
forehead.30 Although this pattern of activation
in the brainstem takes place in a region important in nociceptive
and vascular control (locus ceruleus and dorsal raphe nucleus),
the persistence of the increase in rCBF after relief of symptoms
has been interpreted to suggest the presence of a migraine generator.
In contrast, other information based on molecular genetic investigations
of familial hemiplegic migraine has suggested that altered activation thresholds
within cortical neurons may be important in the initiation of migraine
attacks in some patients.31
++
Approximately 15% of migraineurs experience attacks
in which the headache is preceded by focal neurologic symptoms that
persist for up to 1 hour. These transient symptoms, usually called
the aura, can include transient visual disturbance, unilateral numbness
or weakness, language disturbance, and dizziness and are characterized
by an unexpected onset and a slow expansion of the area affected
by the dysfunction, followed by gradual resolution.
++
Functional neuroimaging in humans during the opening minutes
of spontaneous migraine attacks has shown that decreases in relative
cerebral blood flow occur in areas of occipital cortex. In one case,
the beginning of a migraine attack was fortuitously captured using
PET and 15O-labeled water. A bilateral spreading
area of decreased blood flow was observed that started in visual
associative cortex (Brodmann’s areas 18 and 19) within
a few minutes after the onset of a bioccipital throbbing headache.
The hypoperfusion progressed anteriorly with time across vascular
and anatomic boundaries.32 Although the subject’s
difficulty in focusing on the visual target during a part of the
study has been interpreted as an atypical migraine aura, it is difficult
to draw conclusions about more typical auras, because neither scintillations
nor a scotoma were reported, and because the subject had previously
only had attacks of migraine without aura.
++
More recent studies using functional magnetic resonance imaging
(fMRI) techniques, such as perfusion- and diffusion-weighted imaging,
to study patients during spontaneous migraine visual auras have
shown that during the visual symptoms there are increases in mean
transit time, the amount of time required for a given amount of
blood to move through a set volume of brain parenchyma, of 10% to
54% in the occipital cortex contralateral to the reported
visual field symptoms. These studies have also shown decreases in
both relative cerebral blood flow (15%–53%) and
relative cerebral blood volume (6%–33%)
in that area.33 In between attacks, perfusion weighted
imaging and T2-weighted anatomic images in all of the subjects studied
(eight thus far) were normal. In one subject, in whom multiple perfusion
images were obtained during the same aura, the margin of the perfusion
defect appeared to be more anterior in the second image than in
the first, suggesting a spread reminiscent of Olesen’s
findings.
++
Diffusion-weighted imaging was also performed in subjects during
acute migraine with visual auras. Diffusion-weighted imaging, based
on the mobility of water molecules, reflects the ability of neurons
to maintain normal osmotic membrane gradients. Changes in diffusion-weighted imaging
can indicate a loss of this very basic cell function and are seen
very early in the evolution of ischemic neuronal injury34 and
in the cortex of experimental animals after the induction of spreading
depression.35 Of the subjects for whom diffusion-weighted
images were obtained during visual symptoms, none showed any evidence
of regional hyperintensity to suggest the loss of the ability to
regulate water mobility. No changes in apparent diffusion coefficient
were seen even in areas of occipital lobe that had perfusion decrements
of up to 52%.33 Based on earlier studies in
human stroke, the absence of measurable diffusion abnormalities
suggests that the threshold for ischemia is not crossed during the
migraine aura.34 Negative diffusion data also suggest
that the abnormality that underlies migraine aura in humans, although
possibly analogous, is not identical to spreading depression as
it occurs in experimental animals.
++
In general, recent information from functional neuroimaging tends
to favor a primarily neuronal rather than vascular origin for the
symptoms of migraine aura. The apparent spread across vascular territories
and the moderate blood flow reductions seen in both PET32 and
fMRI33,36 investigations of spontaneous migraine
aura symptoms are more suggestive of primary neuronal dysfunction
than frank ischemia as the basis for the aura. It is interesting
to note that the findings from PET and fMRI (neither of which are
vulnerable to Compton scattering) are consistent with the earlier
observations using 133Xe blood flow techniques.
Whether the characteristics of the neuronal dysfunction prove to
be consistent with a human analogue of spreading depression remains
to be seen.
++
Several lines of evidence are consistent with functional neuroimaging
findings suggesting that dysfunction within the brain is related
to the aura and may provoke head pain. For example, vegetative and
affective prodromal symptoms, such as alteration of mood, appetite,
and fluid balance, may precede the onset of the headache by up to
24 hours. Furthermore, the majority of patients with aura report
that the headache is more severe on the side of the brain hemisphere
to which the aura symptoms localize. In addition, the possibility
that events intrinsic to the cerebral cortex may be capable of activating
meningeal nociceptive neurons is suggested by the fact that seizures
are followed by headache in many patients and that repeated spreading
depressions in experimental animals results in the induction of
c-fos immunoreactivity, (a marker of activation) in second-order
nociceptive neurons within the trigeminal nucleus caudalis.37 Once
depolarized, perivascular and meningeal C fibers transmit nociceptive
information via the trigeminal nerve and upper cervical nerve roots
to areas of pain processing within the trigeminal nucleus caudalis
in the distal medulla and dorsal horn of the upper cervical segments.
++
The gradual intensification and prolongation of headache that
occurs during migraine may be governed through a series of events
that result in peripheral and central sensitization of the trigeminal
system. Information from animal studies indicates that once activated,
C fibers release neuropeptides (i.e., substance P, neurokinin A,
calcitonin gene–related peptide).38 Similar increases
in neuropeptides have been observed during acute migraine attacks
in humans.39 These neuropeptides generate a neurogenic
inflammatory response within the meninges consisting of increased
plasma leakage from meningeal vessels, vasodilation, and activation
of mast cells and endothelial cells.
++
Once set into motion, this process is thought to act to lower
the threshold of the C fibers to further activation and, as a result,
prolong and intensify the headache attack. Drugs known to be effective in
ending a migraine attack such as dihydroergotamine or sumatriptan,
act at serotonin (5-HT1B and 5-HT1D) receptor
subtypes to cause constriction of vascular smooth muscle and to
block the release of neuropeptides that mediate the development
of neurogenic inflammation. Recent animal studies confirm a reduction
in activation thresholds after stimulation of the meningeal pain
system. After chemical stimulation of the meninges, nociceptive
neuronal responses and pain-induced changes in blood pressure were
evoked by much smaller levels of dural mechanical stimulation and
by previously innocuous cutaneous stimulation, indicating the generation
of both central sensitization and cutaneous allodynia.40
+++
Generation of
an Acute Attack
++
Taking into account the information obtained from recent studies,
a possible scenario for the generation of an acute attack in some
forms of migraine is as follows:
++
- 1. Endogenous neurophysiologic events in the neocortex
generate the observed neurologic symptoms of aura and promote the
release of nociceptive substances (e.g., H+ and
K+ ions, arachidonic acid metabolites) from
the neocortex into the interstitial space.
- 2. Within the Virchow-Robin spaces, the released substances
accumulate to levels sufficient to activate or sensitize trigeminovascular
fibers that surround pial vessels supplying the draining neocortex.
- 3. Substances that discharge or sensitize small, unmyelinated
fibers that transmit pain accumulate in proximity to trigeminovascular
fibers and may possibly provide the trigger for headache or sensitize
perivascular afferents to blood-borne or other as yet unidentified
factors. The headache latency (20 to 40 minutes) may reflect the
time needed for extracellular levels to reach a threshold for depolarization.
- 4. Upon activation, the trigeminal nociceptive neurons transmit
the nociceptive information through the trigeminal ganglia to synapse
on second-order neurons within the trigeminal nucleus caudalis.
- 5. Numerous projections from the trigeminal nucleus caudalis
then transmit the nociceptive information to various brain areas
that underlie various aspects of pain (see the earlier discussion
of “Anatomy of Head Pain”).
++
In this scenario, the brain acts as a transducer, interfacing
with the environment. Triggering events, such as those associated
with emotional stress, glaring lights, or interrupted sleep, modulate
activity within brain regions physically contiguous to the meningeal
vessels innervated by the trigeminal nerve. In susceptible individuals,
these events may be sufficient to initiate neurophysiologic events
leading to chemical activation of meningeal fibers. The photophobia,
nausea, and vomiting are probably not specific to migraine but are
related to meningeal irritation, as similar symptoms are seen with
infection or when blood enters the subarachnoid space. This proposed cascade
provides a pathogenetic framework for further investigation of migraine
and is based on currently understood principles of neurobiology
and the physiology of pain. However, the details will need revision
as new data emerge from experimental studies in humans and animals.