One Hundred Years of Migraine Research: Major Clinical and Scientific Observations From 1910 to 2010

Abstract

(Headache 2011;51:752-778)

Pain research, and headache research in particular, during the 20th century, has generated an enormous volume of literature promulgating theories, questions, and temporary answers. This narrative review describes the most important events in the history of migraine research between 1910 and 2010. Based on the standard textbooks of headache: Wolff's Headache (1948 and 1963) and The Headaches (1993, 2000, and 2006) topics were selected for a historical review. Most notably these included: isolation and clinical introduction of ergotamine (1918); further establishment of vasodilation in migraine and the constrictive action of ergotamine (1938); identification of pain-sensitive structures in the head (1941); Lashley's description of spreading scotoma (1941); cortical spreading depression (CSD) of Leão (1944); serotonin and the introduction of methysergide (1959); spreading oligemia in migraine with aura (1981); oligemia in the wake of CSD in rats (1982); neurogenic inflammation theory of migraine (1987); a new headache classification (1988); the discovery of sumatriptan (1988); migraine and calcitonin gene-related peptide (1990); the brainstem “migraine generator” and PET studies (1995); migraine as a channelopathy, including research from the genetic perspective (1996); and finally, meningeal sensitization, central sensitization, and allodynia (1996). Pathophysiological ideas have evolved within a limited number of paradigms, notably the vascular, neurogenic, neurotransmitter, and genetic/molecular biological paradigm. The application of various new technologies played an important role within these paradigms, in particular neurosurgical techniques, EEG, methods to measure cerebral blood flow, PET imaging, clinical epidemiological, genetic, and molecular biological methods, the latter putting migraine (at least hemiplegic migraine) within a completely new classification of diseases.

Physicians tend to plot the history of medicine as a series of successes, through which a logical line can be drawn from the past to the present that may even be extrapolated to the future, ultimately leading to the complete understanding of natural phenomena. But when we consider how science progresses, we realize that it is a process of trial and error. Furthermore, with respect to this process, epistemological insights change over time. The scientific truth of today is the lie of tomorrow and as the Dutch historian Huizinga said, history is the “intellectual form in which a culture renders account of its past.” Each generation will consider the history of a certain episode from another perspective. Pain research during the 20th century, and headache research in particular, has resulted in an enormous number of papers and books full of theories, questions, and answers. The number of effective therapies has increased even if some of the hypotheses upon which they were based have since been superseded. Cognizant of shifts in our understanding, we endeavor to describe what we consider to be the most important events in the history of migraine research between 1910 and 2010. We also contextualize these events within contemporary medical research. Almost certainly, by 2050 our interpretation of these events will have changed again. Consequently, documenting today's perspective represents a useful exercise.

METHODS

Our approach to interpreting the important events in headache research involved selected papers that we considered influential studies during our study period (1910 to 2010). The topics were not generated randomly, but selected by a search through 2 major textbooks on headache: notably Wolff's monographs on headache (1948 and 1963) as midcentury representative (Wolff Headache and Other Head Pain, 1948, 1963^1,2) and Olesen et al's multi-authored textbook (The Headaches^3-5) as the representative review of knowledge at the turn of the 20th to the 21st century. We emphasized pathophysiological ideas and treatment options evolving from it. Finally, we searched Google Scholar and ISI Web of Knowledge for citations of the selected papers.

RESULTS

We identified 15 major areas in the study of migraine (see Table 1) and the relevant papers^6-20 and related papers are reviewed in the following.

Table 1.—. Major Clinical and Scientific Observations in Migraine Research From 1900 to 2000

Subjects	References
Isolation and clinical introduction of ergotamine	Stoll, 1918^6
Further establishing the vasodilation in migraine and the constrictive action of ergotamine	Graham and Wolff, 1938^7
Pain-sensitive structures in the head	Ray and Wolff, 1940^8
Lashley's description of spreading scotoma	Lashley, 1941^9
Cortical spreading depression (CSD) of Leão	Leão, 1944^10
Serotonin and the introduction of methysergide	Sicuteri, 1959^11
Spreading oligemia in migraine with aura	Olesen et al, 1981^12
Oligemia in the wake of CSD in rats	Lauritzen et al, 1982^13
Neurogenic inflammation theory of migraine	Moskowitz, 1984^14
A new headache classification	Classification Committee of the International Headache Society, 1988^15
A new drug for migraine – the discovery of sumatriptan	Humphrey et al, 1988^16
Migraine and calcitonin gene-related peptide	Goadsby and Edvinsson, 1990^17
The brainstem “migraine generator”– PET studies in migraine	Weiller et al, 1995^18
Identification of the gene for familial hemiplegic migraine	Ophoff et al, 1996^19
Meningeal sensitization, central sensitization and allodynia	Burstein et al, 1996^20

Ergotamine (1918-1938).— One of the most important milestones in the early 20th century was the isolation and clinical introduction of ergotamine. Woakes had recommended ergot for the treatment of migraine in 1868.²¹ Sir Henry Dale discovered that the liquid extract blocked the effects of stimulation of the sympathetic nerves.^22,23 It would appear later that this was a question of dosage, lower dosages being vasocontrictive. Stoll isolated ergotamine from ergot⁶ and Rothlin believed that its adrenolytic properties would counteract the sympathicotonic effects in migraine (evidenced by a pale face in some patients) by using this vasodilator.²⁴ Maier applied it to 80 patients with “sympathicotonic conditions” (migraine, types of epilepsy, psychiatric diseases, urticaria, and Basedow's disease).²⁵ Using placebo controls, Trautmann found the drug effective.²⁶ Tzanck²⁷ presented positive results and suggested to use ergotamine in “équivalents gastriques de la migraine,” including asthma, cyclic vomiting, herpes, postlumbar puncture headache, and sea sickness. He believed to treat the sympathicotonic state, referring to Du Bois-Reymond²⁸ from 1860,^23,29 and published data on 101 patients 3 years later.³⁰

Ergotamine was introduced in the USA^31-33 and intravenous ergotamine proved effective in 90% of 109 patients.³⁴ Blood pressure changes and uterine contractions were noted to begin almost at once but relief of headache not before nearly 1 hour, pointing to the time-effect curve for the effect on arteries in man.³⁵ This is in contrast to some of the findings of Graham and Wolff (Fig. 1⁷, vide infra). Outstanding effects were published³⁶ and parenteral ergotamine appeared more effective than the oral form.³⁷

Figure 1.—. One of the famous pictures from the paper by Graham and Wolff, 1938. Relation between pulsation amplitude of the temporal artery and the headache intensity following the administration of ergotamine tartrate. Representative records of the photographic record are inserted. The initial temporal artery amplitude was taken as 99%.⁷

The introduction of ergotamine and the doubts about the existing pathophysiological ideas on migraine inspired Graham and Wolff, who studied both the external carotid vessels, directly by measuring the amplitude of pulsations following ergotamine injections, and the intracranial vessels, indirectly, by measuring cerebrospinal fluid (CSF) pulsation in the lumbar subarachnoid space. There was a close relationship between the decrease in amplitude and the decline of headache intensity, resulting in one of the most important figures in migraine research of the 20th century, and determining further research of the vascular hypothesis (Fig. 1⁷). A relationship with the CSF pulsations, supposedly reflecting the amplitude of the intracranial arteries, or CSF pressure, was not observed. They concluded that “the most acceptable explanation of the headache-ending effect is that cranial arterial walls which are painfully stretched and dilated are caused to narrow through the vasoconstrictor action of ergot” and thereby refuted the sympathicotonic theories of the 1920s. For many years ergotamine and its derivative dihydroergotamine were the only specific antimigraine drugs. A more recent European consensus found it the drug of choice in a limited number of migraine sufferers who have infrequent or long duration headaches.³⁸

Pain-Sensitive Structures in the Head (1940).— The study of pain-sensitive structures by Ray and Wolff in the 1930s was of great importance but certainly not new. It was mentioned in many of the ancient texts on headache, including Van Beverwijck's Treasure of Unhealthiness of 1642.³⁹ But Ray and Wolff were able to use new techniques for their observations, notably neurosurgical procedures carried out with local anesthesia during which communication with the patients about their experiences was possible. This paper⁸ is still a mainstay of migraine literature and remains frequently cited (see Table 2). For the study concerned, they selected 30 “intelligent and cooperative” patients “free of apprehension and of preoccupation with pain, so that a minimal amount of local and general analgesia was required,” undergoing neurosurgical procedures. Several extra- and intracranial structures were studied by faradic stimulation, including the scalp, galea, fascia, muscles, arteries, veins, and sinuses in 150 observations in 30 subjects. The figures drawn from all the experiments are instructive with respect to the areas where the (referred) pain was felt. An example is shown in Figure 2 in which stimulation of the middle meningeal artery resulted in temporal pain, Also, from other pain sensitive structures such as proximal cerebral arteries, larger intracranial, veins, and part of dura, there was a distinct localization of the pain. Several conclusions were drawn. Extracranially, most tissues are sensitive, the arteries in particular. Intracranially, the great venous sinuses and “venous tributaries from the surface of the brain,” as well as parts of the dura at the skull base, the dural arteries and the “cerebral arteries at the base of the brain,” are sensitive to pain. Structures not sensitive to pain include the skull, the brain parenchyma, most of the dura covering it, most of the pia-arachnoid, the ependymal lining of the ventricles, and the choroid plexuses. Of further importance was the observation that “stimulation of the pain-sensitive intracranial structures on or above the superior surface of the tentorium cerebelli resulted in pain in various regions in front of a line drawn vertically from the ears across the top of the head,” the pathways running through the trigeminal nerve. Stimulation on or below the inferior surface of the tentorium resulted in pain in various regions behind this line, the pathways running through the glossopharyngeal and vagus nerve, as well as the 3 upper cervical roots.⁸

Table 2.—. Highlight Papers Since the 1930s (Before That, No Citations Available in Databases) and Numbers of Citations in Google Scholar (and ISI Web of Knowledge in Parentheses) by June 2009† (for References, See Table 1)

Paper	Subject	Citations
† Citation Index was not used because we could not find Leão(1944).^10
Graham and Wolff from 1938	Pulsation of temporal artery during migraine	256(283)
Ray and Wolff from 1941	Pain sensitive structure in the head	482(449)
Lashley from 1941	Spread of fortification spectra	234(246)
Leão from 1944	Cortical spreading depression(CSD)	1129(1287)
Sicuteri from 1959	Role of methysergide and serotonin in migraine	83(?)
Olesen et al from 1981	Spreading oligemia in migraine with aura	411(482)
Lauritzen et al from 1982	Oligenia in the wake of CSD in rats	80(132)
Moskowitz from 1984	Neural activation releases vasoactive neurotransmittors	494(565)
Headache classification from 1988	The first headache classification with operational diagnostic criteria	301(?)
Humphrey et al 1988	The introduction of sumatriptan	225(381)
Goadsby et al from 1990	Calcitonin gene-related peptide was increased in external jugular vein during migraine attacks	471(486)
Weiller et al from 1995	PET showed activation of brain stem during migraine attacks	675(514)
Ophoff et al from 1996	Migraine as a channelopathy	1215(1132)
Burstein et al from 2000	Cutaneous allodynia suggesting central sensitization during migraine attacks	393(270)

Figure 2.—. One of the many figures from the paper by Ray and Wolff. Stimulation of the middle menigeal artery caused pain in the ipsilateral temporal region.⁸

In retrospect, we need to recognize that Ray and Wolff used localized short-lasting faradic stimulation, but both spatial and temporal summations are integral mechanisms of pain, particularly in persistent pain conditions.^40-42 While focal and short-lasting stimulation of the dura mater or of a small blood vessel in the pia mater are not painful, it is likely that long lasting stimulation and/or stimulation of a large area of the dura mater or the pia may be painful. Supporting this possibility are clinical documentations of extreme pain during meningitis and subarachnoid hemorrhage.⁴³

Lashley's Description of Visual Auras (1941).— Although scientific study of the migraine aura had been undertaken in the 19th century by scientists who suffered from migraine,^44-47 a study that significantly furthered the understanding of the phenomenon was that carried out by the American psychologist Karl Spencer Lashley (1890-1958), a professor of psychology at Harvard University.⁹ The study of his own visual auras (which numbered upward of 100) was described in a remarkable and influential paper in 1941. He referred to 2 extensive previous reviews by Richter⁴⁸ and by the Berlin migraine sufferer/neuropsychiatrist Friedrich Jolly (1844-1904), who like Airy⁴⁴ had noticed that flickerscotomas as in migraine are a frequent nuisance of the class of scientists.⁴⁹ Lashley noticed that 2 characteristics had not been reported previously: the maintenance of the characteristic shape of the scotoma during its drift across the visual field and the “completion of figure.”“Over a period of years I have had opportunity to observe and map a large number of such scotomas, uncomplicated by any other symptoms of migraine,” observed Lashley. He mapped the figures he observed in space and time (Fig. 3⁹). He saw that the form of the figures is usually maintained during the evolution of the aura and “when there are fortification figures, these also maintain their characteristic pattern in each part of the area.” He suggested that “an inhibitory process, in the case of the blind areas, or an excitatory process, in the case of scintillations, is initiated in one part of the visual cortex and spreads over an additional area.” Thus, distinguishing the excitatory from the inhibitory part of the aura, he realized that during the spreading of the process, “activity at the point where it is initiated is extinguished, and the process of extinction also spreads over the same area at about the same rate as does the active process.” Lashley was able to determine the rate of spread. “Ten to twelve minutes is required for spread of the outer margin from the region of the macula to the blindspot of the homolateral eye” and the total time for the spread from the macular to the temporal area was what we also hear from our patients: 20 minutes. The anteroposterior length of the striate area being about 67 mm, he concluded that the “wave of intense excitation is propagated at a rate of 3 mm/minute or less across the visual cortex” and that “the wave is followed by complete inhibition of activity, with recovery progressing at the same rate,” adding that sometimes “the inhibition spreads without the preceding excitatory wave.” Later Lashley's theories had great impact, not least because of the description of Leão's cortical spreading depression (CSD) in 1944,¹⁰ 3 years after his paper was published in 1941.⁹

Figure 3.—. From Karl Lashley's 1941 paper describing his own auras. “Over a period of years I have had opportunity to observe and map a large number of such scotomas, uncomplicated by any other symptoms of migraine.” He mapped the figures he observed in space and time and concluded that the cortical process spread with a velocity of 3 mm per minute.⁹

Cortical Spreading Depression of Leão (1944).— After the first study of the EEG in 1929 by Hans Berger (1873-1941) and its popularization by Nobel Prize winner Edgar Adrian (1889-1977) in the 1930s, EEG studies became widely available. CSD was discovered in 1943 by Aristides Leão (1914-93), a Brazilian neurophysiologist, during his PhD fellowship at the physiology department of Harvard University. His results were first published in 1944.¹⁰ At the time Leão had been conducting research on “experimental epilepsy” in the cerebral cortex under the supervision of Hallowell Davis (1896-1992) and Arturo Rosenblueth (1900-70).⁵⁰ In cooperation with Rosenblueth he studied electric activity in a rabbit brain under general anesthesia. Following electrical stimulation a most unexpected and contradictory result was observed: “the activity of the nearest pair of electrodes did not increase, but ceased almost entirely.” Davis was called in for consultation and said “nothing resembles a new phenomenon as much as a good artifact.”⁵⁰ The response, however, was reasonably reproducible.¹⁰ It consisted of a marked, enduring, reduction of electrical activity, a reduction which appears first at the region that has been stimulated, and spreads out from that location in all directions, involving successively more and more distant parts of the cerebral cortex (Fig. 4). The recovery usually took 5-10 minutes. In a second paper, Leão described a wave of marked dilatation of the pial vessels traveling over the cerebral hemispheres concomitant with the CSD.⁵¹ In a third paper, it was demonstrated that CSD was not inhibited by anoxia.⁵² The paper proposed that CSD might be related to migraine with aura because of the slow development of scotomata and sensory symptoms of migraine aura.^10,52 It should be noted, however, that the authors were unaware of Lashley's 1941 description.⁵⁰ Interestingly, Leão did not attempt to calculate the speed of CSD in these 3 papers.^10,51,52 It was later calculated to be 3 mm/minute.⁵³ Milner in 1958 in a short communication drew attention to the similarity of the findings of Leão and Lashley.⁵⁴

Figure 4.—. Well-known EEG graph picturing cortical spreading depression (CSD) from Aristides Leão's 1944 paper.¹⁰ Reduction of electrical activity appearing first at the region that has been stimulated and spreading out from there in all directions, involving successively more and more distant parts of the cerebral cortex.

The relationship between CSD and migraine was first studied in the 1980s, when spreading oligemia was observed during migraine with aura¹² (vide infra). The literature on CSD is significant in its scale and beyond the scope of this review. For recent updates, see the studies by Smith et al and Charles and Brennan.^55,56

Serotonin and the Introduction of Methysergide (1959).— Between 1948 and 1953, serotonin, a serum (“sero”) vasoconstrictor (“tonin”) factor, was identified, isolated, and synthesized. In the 1950s and 1960s, its role in migraine was gradually established by Wolff et al.⁵⁷ Serotonin was one of the agents they examined and by perivascular injection, they were able to produce migraine-like symptoms.⁵⁸ The search for an effective 5-HT antagonist led to the synthesis of methysergide, derived from LSD25 that is an effective agent with this respect, but hallucinogenic. In 1959, methysergide was introduced in the clinic as a drug for the preventive treatment of migraine by Federico Sicuteri, an Italian neurologist.¹¹ As migraine and cluster headache were both considered “vasodilating headaches,” both kinds of patients were entered in Sicuteri's study and he considered the results most promising. Doing further research on serotonin and migraine, he found increased excretion of 5-HIAA during migraine attacks.⁵⁹ Two well-conducted randomized controlled trials (RCTs) in the mid-1960s^60,61 clearly showed the superiority of methysergide to placebo.

Five years after its introduction in the clinic, however, the first serious side effects, notably retroperitoneal fibrosis,⁶² were published. The frequency was estimated at 1/1000. The frequency was later reestimated at 1/5000.⁶³ Cardiac and pulmonary fibrosis was described shortly thereafter,⁶⁴ but in lower frequencies. It is still used as a fourth choice prophylactic drug in migraine and cluster headache, the administration of which is discontinued for 2 months every half a year.

Interestingly, the discovery of methysergide provided an incentive for the development of sumatriptan, when evidence was found for an “atypical” 5-HT receptor in the carotid bed of pigs that was later identified as a 5-HT_1B receptor.⁶⁵ Methysergide appeared to have a partial agonist action upon this receptor, whereas it blocks 5-HT₂ receptor.⁶⁶ Therefore, methysergide not only is an effective prophylactic drug, but also played an important role as a 5-HT antagonist and partial agonist in pharmacological studies.

Spreading Oligemia of Cerebral Blood Flow (1981).— The vascular theory with respect to the pathophysiology of migraine led researchers to study regional cerebral blood flow (rCBF).⁶⁷ The prerequisite for a precise characterization of rCBF during migraine was the development of a multichannel imaging system with intracarotid injection of Xenon-133 and 254 detectors, which resulted in a spatial resolution of 1 cm by Lassen et al.⁶⁸ Six patients with migraine with aura were followed with serial measurements of rCBF with the intracarotid Xenon-133 method from the normal state into the aura phase, and in 3 cases into the headache phase¹² (see Fig. 5). During the aura phase all patients developed rCBF reduction (oligemia), which only in one case approached critical values. Oligemia gradually spread anteriorly in the course of 15 to 45 minutes (Fig. 5). In 4 cases severe headache occurred concomitantly with oligemia. The paper was concluded by stating that “the results indicate that the vasospastic model of the migraine attack is too simplistic. Alteration in neuronal function, in the blood–brain barrier (BBB), or in some other brain process is more likely to be the primary event of the attack.”¹² CSD of Leão was considered (Lauritzen, personal communication, 2008), but as noted above, at the time, CSD was known to be associated with hyperemia (vide supra)^51,69 and therefore contradicted the assumption of CSD being the primary neuronal process underlying the spreading oligemia.

Figure 5.—. Serial measurements of regional cerebral blood flow (rCBF) with the intracarotid Xenon-133 method from the normal state into the aura phase (OR into the headache phase).¹² The attack was initiated by focal hyperemia and during the aura phase the patients developed rCBF reduction (oligemia). Oligemia gradually spread anteriorly in the course of 15 to 45 minutes.

The migraine patients in Olesen's study¹² belonged to a series of approximately 250 patients undergoing carotid arteriography for various diagnostic reasons. The carotid catheterization with a catheter placed using the Seldinger technique and angiography most likely induced the migraine aura. The paper by Olesen et al¹² and shortly after a paper by Lauritzen et al,⁷⁰ both on rCBF in migraine with aura, and their follow-up investigations of rCBF and CSD,^70-74 had an enormous impact on the concept of migraine as a brain disease.

In contrast, in migraine without aura, a normal rCBF was found.⁷⁵ In 1990, 10 years of rCBF studies in migraine were summarized.⁷⁶ Results of rCBF in 63 migraine with aura patients were analyzed. Twenty patients had been investigated with the intracarotid technique⁶⁸ and 42 patients with single photon emission computed tomography (SPECT).⁷⁷ Focally reduced rCBF was often observed before the patients experienced aura symptoms.⁷⁶ With SPECT the later phase of spontaneous migraine attacks could be studied. During the headache phase, rCBF gradually changed from abnormally low to abnormally high without apparent changes in headache.^74,76 A summary of these findings is shown in Figure 6.⁷⁶

Figure 6.—. Schematic illustration in migraine with aura of the temporal relation between angiography (time 0 hour), hypoperfusion, aura, headache, hyperperfusion, disappearance of headache, and disappearance of hyperperfusion. Note the discrepancy between the headache and cerebral blood flow changes.⁷⁵

In 2001 Hadjikhani et al in a landmark study, investigated 3 patients during visual aura using functional magnetic resonance imaging (fMRI).⁷⁷ One patient with an exercise-induced aura showed a focal increase in BOLD signal (probably reflecting vasodilation), developed within extrastriate cortex and this BOLD change progressed slowly (3 mm/minute). Then the BOLD signal diminished (possible reflecting vasoconstriction) (Fig. 7). This indicated that an electrophysiological event such as CSD generated the aura in the visual cortex.⁷⁷ One patient who had an atypical migraine attack showed bilateral spreading hypoperfusion in a PET study followed by migraine headache.⁷⁸ The findings were only minimally influenced by scattered radiation (see above) and the study documented beyond reasonable doubt that spreading hypoperfusion is a real phenomenon.⁷⁹

Figure 7.—. Spreading suppression of cortical activation during exercise-induced visual migraine aura.⁷⁷ (A) A drawing showing the progression over 20 minutes of the scintillations and the visual field defect affecting the left hemifield, as described by the patient. (B) A reconstruction of the same patient's brain, based on anatomical MR data. The posterior medial aspect of the occipital lobe is shown in an inflated cortex format. A marked perturbation or change in the BOLD signal was found. It developed sequentially along consecutive regions of calcarine cortex and his perturbation in the BOLD signal was identical from voxel-to-voxel along calcarine cortex, differing only with respect to the time of onset, beginning posteriorly and spreading anteriorly. It was inferred from this study that migraine aura is accompanied by a propagating brain event, which is retinotopically related to the visual perception (moving from central to peripheral visual fields). The changes in BOLD signal was interpreted as first, an increase in blood flow lasting a few minutes, followed by a longer lasting decrease in blood flow that dropped below baseline levels. The authors concluded that this sequence of events is similar to what is observed during cortical spreading depression and documented repeatedly in rodent or cat cortex.⁷⁷

In a recent PET study in spontaneous migraine without aura attacks investigated within 4 hours of debut, bilateral occipital hypoperfusion was observed.⁸⁰ This finding questioned the separation of migraine with and without aura on pathophysiological grounds. It should be noted that in 2 other PET studies,^18,81 there was no posterior cortical hypoperfusion in migraine without aura. Further confirmation of this finding is still needed.⁸⁰

Four patients were investigated in 1998 with perfusion- and diffusion-weighted magnetic resonance imaging during spontaneous visual aura.⁸² A 16-53% decrease in relative CBF was observed in the gray matter of the occipital cortex contralateral to the affected visual field. No changes in the apparent diffusion coefficient were observed.⁸² Eight migraine patients developed atypical visual changes and/or headache during visual stimulation.⁸³ In 5 patients there was an initial activation pattern as measured with MRI-BOLD.⁸³ This was followed by a suppression of activity and the area of suppression progressed across the occipital cortex at a rate of 3-6 mm per minute.⁸³ In a later series of cases from the same group, migraine attacks triggered by visual stimuli were found in 75% of cases to have increased MRI-BOLD signal in the red nucleus and substantia nigra prior to changes in the occipital cortex, implicating these structures in migraine both with and without aura.⁸ For a more detailed overview of the history of rCBF and migraine, see the study by Tfelt-Hansen.⁸⁴

Oligemia in the Wake of Cortical Spreading Depression (1982).— Inspired by the rCBF results in migraine with aura¹² and using quantitative autoradiography, in 1982 Lauritzen et al investigated CBF in rats during, and in the wake of, CSD.¹³ As shown in Figure 8, cortical blood flow increased 218% during the CSD wave, but, more importantly, it decreased 15% to 27% after the hyperemia and for more than 1 hour after CSD. The changes in blood flow were largely limited to the cerebral cortex. This was the first time that oligemia was observed in connection with CSD and the authors speculated that “the spreading oligemia of migraine with aura may be a phenomenon physiologically related to the finding of oligemia after CSD.”¹³ The finding of initial hyperemia followed by oligemia in connection with CSD was later confirmed in anesthetized cats⁸⁵ and awake and freely moving rats.⁸⁶

Figure 8.—. Persistent oligemia in wake of cortical spreading depression (CSD). Cortical blood flow increased 218% during the CSD wave, decreased 15% to 27% after the hyperemia and for more than 1 hour after CSD. The changes in blood flow were largely limited to the cerebral cortex.¹³

A next step was measuring rCBF in migraine patients undergoing carotid angiogram for diagnostic purposes and the carotid technique again induced migraine with aura.⁷¹ rCBF was measured repeatedly at short intervals in order to document the slow spread of hypoperfusion. A wave of reduced rCBF originating in the posterior part of the brain slowly progressed anteriorly with a speed of 2 mm per minute.⁷¹ Four of 13 patients developed headache during the rCBF study at the time of global oligemia. It was suggested that focal symptoms and rCBF changes might be secondary to CSD.⁷¹ In the second part of the study, cerebrovascular reactivity to voluntary hyperventilation, moderate hypertension, and physiological activation were studied.⁷⁰ During attacks the carbon dioxide reactivity (change in rCBF per mmHg change in PaCO₂) was decreased to 3% in the oligemic regions compared with 6% in the normally perfused brain. Blood pressure was normal in all brain regions. Similarly, the CO₂ response after CSD in rats was impaired whereas autoregulation was preserved.⁸⁷ The similarities of spreading oligemia of rCBF during migraine aura and CSD strongly supported the hypothesis that the migraine aura is caused by CSD.⁸⁷

In one study in rats CSD caused a long-lasting blood flow enhancement selectively within the middle meningeal artery.⁸⁸ In addition, CSD provoked plasma protein leakage within the dura mater. The results provided a neural mechanism, dependent on trigeminal and parasympatic activation, by which extracerebral cephalic blood flow couples to CSD and it was suggested that a similar mechanism in man explains the headache in migraine with aura.⁸⁸

Cortical spreading depression may alter BBB permeability by activating brain matrix metalloproteinases (MMPs).⁸⁹ Beginning after 3-6 hours, MMP-9 levels increased within cortex ipsilateral to CSD reaching a maximum at 24 hours and persisting for at least 48 hours. At 3 hours after CSD, plasma protein leakage and brain edema developed contemporaneously. It was concluded that CSD may cause MMP upregulation and increased vascular permeability in migraine, stroke, and trauma.⁹⁰ For a discussion of the BBB in migraine, see the study by Edvinsson and Tfelt-Hansen.⁹⁰ In a study from 2004, a CACNA1a knock-in mouse was used as a model of familial hemiplegic migraine (FHM). It showed increased susceptibility to CSD⁹¹ indicating that CSD could be involved in FHM1.

In a study from 2006 the effect of migraine prophylactic drugs on CSD was studied.⁹² Rats were treated either acutely or chronically over weeks and months with topiramate, valproate, propranolol, amitriptyline, or methysergide. Chronic daily administration of migraine prophylactic drugs dose-dependently suppressed CSD frequency by 40% to 80% and increased the cathodal stimulation threshold, whereas acute treatment was ineffective. The findings suggested that CSD provides a common therapeutic target for widely prescribed migraine prophylactic drugs.

In 14 patients undergoing neurosurgery after head injury or intracranial hemorrhage, electrocorticographic (ECoG) electrodes were placed near foci of damaged cortical tissue.⁹³ Transient episodes of depressed ECoG activity that spread across the cortex at rates of 0.6 and 5 mm/minute were observed in 5 patients. The rate of progression suggested a CSD. In a further similar study, evidence for CSD or peri-infarct depolarization (PID) was recorded in 50% of patients.⁹⁴ In a third study, 50% of 63 patients had a CSD after acute cerebral injury.⁹⁵ Similarly, CSD and/or PID occurred in all but 2 of 16 patients with large middle cerebral artery infarction.⁹⁶ These studies from 2002 to 2008^93-95,97 demonstrated beyond doubt that CSD can occur in the human brain.

Neurogenic Inflammation Theory of Migraine (1984).— In a hypothesis paper from 1979, it was suggested that the abnormal release of substance P, found in the trigeminal nerve and released by depolarization of as yet unidentified peptides or other neurotransmitters from the fifth cranial nerve, may explain both the hemicranial pain and the vasodilation, which are characteristic of the headache of migraine.⁹⁶ The investigators were prompted by the clinical characteristics of migraine attacks to focus attention, as was carried out by others more than 30 years earlier,^2,8 on the importance of the trigeminal system in the pathogenesis of headache.⁹⁸ By applying horseradish peroxidase to the middle cerebral artery in cats, cell bodies containing the enzyme marker were located among a cluster of cells in the ipsilateral trigeminal ganglion. Hereby, a neuroanatomical pathway between cerebral arteries and brain was demonstrated.⁹⁹

The next step was to determine whether substance P, vide supra, was present in the trigeminovascular fibers. Levels of substance P-like immunoreactivity were examined using pia arachnoid and its attendant blood vessels in vitro.¹⁴ They found that substance P was present in many animal species' pia plus arachnoid (containing pial arteries), or in pial arteries alone.^14,100,101 In an extensive and influential review (see Table 2), it was suggested that neural activation releases vasoactive neurotransmitters from their afferent processes, which in turn provokes inflammatory changes in peripheral target tissues (in this instance, cerebral blood vessels).¹⁴ In addition, the data suggested that the release of substance P (a vasodilator) from sensory fibers is important in mediating changes in vessel permeability. The finding in animals explains the unilateral distribution of migraine pain and could challenge the concept that the pain of vascular headache is due to dilating blood vessels. Finally, it was suggested that there is a sterile inflammation of cranial blood vessels during migraine attacks. Moskowitz concluded that “the relationship of trigeminovascular fibers to the pathogenesis of vascular head pain sheds light on possible mechanisms of migraine and other central nervous system conditions associated with headache and inflammation.”¹⁴

Studies along this line were carried out in the subsequent year and in 1987 a model of intracranial neurogenic inflammation was presented for use in rat.¹⁰² Electrical stimulation of the trigeminal ganglion resulted in ipsilateral increase of tracers in the dura. In contrast, there was no extravasation in the brain.¹⁰² The neurogenic inflammation model was then used to demonstrate that ergot alkaloids, ergotamine, and dihydroergotamine, inhibited plasma extravasation and it was suggested “that the therapeutic effects of ergots in vascular headaches may result from peripheral blockade of small fiber (C or A-delta)-dependent neurogenic inflammation within the dura.¹⁰³ Similar results were obtained for indomethacin and aspirin¹⁰⁴ and sumatriptan.¹⁰⁵ It was suggested that the effect was mediated by 5-HT1B/1D receptors located on sensory trigeminal neurons.^106,107

Up to now all effective, acute antimigraine drugs have indeed been proven to inhibit neurogenic protein extravasation (NPE).^105,107,108 However, inhibition of NPE was not predictive for the antimigraine effect of all investigated new drug-groups. Thus, in 5 placebo-controlled clinical studies, drugs (eg, endothelin and NK-1 receptor antagonists) with potent inhibitory effect on dural NPE were not effective for acute migraine treatment.^109-113 Because specific NPE inhibitors are without an effect in migraine, it has been difficult to find a pivotal role for dural neurogenic inflammation in migraine. It was shown, however, that valproate, which is effective in migraine prophylaxis, blocked plasma extravasion in the meninges.¹¹⁴

In one study glyceryl trinitrate (GTN) infusion, which provokes headache in healthy volunteers¹¹⁵ and in addition delayed migraine in migraineurs,¹¹⁶ caused a delayed inflammation in rat meninges,¹¹⁷ suggesting that a similar response develops in human meninges after GTN challenge.

The Need for a Headache Classification (1988).— Although the need for a unified headache classification had been mentioned previously in the 20th century, it lasted until 1962 when an Ad Hoc Committee classified headaches, but the use of words like “usually” and “commonly” made the definition widely open for personal interpretation.¹¹⁸ Thus, migraine was defined as:

Recurrent attacks of headache, widely varied in intensity, frequency, and duration. The attacks are commonly unilateral in onset; are usually associated with anorexia, and sometimes with nausea and vomiting; and some are preceded by, or associated with, conspicuous sensory, motor, and mood disturbances; and are often familial.¹¹⁹

A major breakthrough in headache research was the work of the Headache Classification Committee (headed by the Danish researcher Jes Olesen) of the International Headache Society (founded in 1981), resulting in the first extensive headache classification with operational diagnostic criteria in 1988.¹⁵ Headache was classified into 14 groups with 4 primary headache groups, including migraine, tension-type headache, cluster headache, and other primary headaches. The other 10 groups concerned secondary headaches. Operational diagnostic criteria were described for each entity like migraine without aura and migraine with aura (see Tables 3 and 4). This classification with operational diagnostics ensured that scientific research could be performed globally in comparable patients' populations. Similar to the Ad Hoc Classification from 1962, migraine was subdivided, now on scientific grounds, into migraine without aura, formerly “common migraine,” and migraine with aura, formerly “classic migraine.” As discussed above, pathophysiologically these 2 forms differ mainly in rCBF during attacks: in migraine with aura there is a decrease in rCBF, a spreading oligemia, during aura and into the headache phase,^12,76 whereas rCBF is unchanged during migraine without aura.⁷⁵ These differences in rCBF were among the most convincing arguments for the separation of the 2 migraine forms and against a continuum model of migraine. The classification was first used on a large scale in the extensive clinical trials program of sumatriptan¹²⁰ and later in the trial program of the other triptans.^121,122

Table 3.—. International Headache Society Diagnostic Criteria for Migraine Without Aura (Classification Committee of the International Headache Society, 1988¹⁵)

A.At least 5 attacks fulfilling B through D

B. Headache lasting 4 to 72 hours (untreated or unsuccessfully treated)

C. Headache has at least 2 of the following characteristics

1. Unilateral location

2. Pulsating quality

3. Moderate or severe pain intensity (inhibits or prohibits daily activity)

4. Aggravation by walking stairs or similar routine activity

D. During headache at least one of the following

1. Nausea and/or vomiting

2. Photophobia and phonophobia

E. At least one of the following

1. History and/or physical and/or neurological examination do not suggest one of the disorders listed in groups 5-11

2. History and/or physical and/or neurological examinations do suggest such disorders, but is ruled out by appropriate investigations

3. Such disorder is present, but migraine attacks do not occur for the first time in close temporal relation to the disorder

Table 4.—. International Headache Society Diagnostic Criteria for Migraine With Aura (Classification Committee of the International Headache Society, 1988¹⁵)

A.At least 2 attacks fulfilling B

B. At least 3 of the following 4 characteristics

1. One or more reversible aura symptoms indicating focal cerebral cortical and/or brainstem dysfunction

2. At least one aura symptom develops gradually over more than 4 minutes or 2 or more symptoms occur in succession

3. No aura symptom last more than 60 minutes. If more than aura symptom is present, accepted duration is proportionally increased

4. Headache follows aura with a free interval of less than 60 minutes (it may also begin before or simultaneously with the aura)

E. At least one of the following

1. History and/or physical and/or neurological examination do not suggest one of the disorders listed in groups 5-11

2. History and/or physical and/or neurological examinations do suggest such disorders, but is ruled out by appropriate investigations

3. Such disorder is present, but migraine attacks do not occur for the first time in close temporal relation to the disorder

A second revised version was published in 2004 (International Classification of Headache Disorders-II).¹²³ The major changes were in the migraine with aura group where typical aura could be followed by either migraine headache or just headache. Sporadic hemiplegic migraine was recognized as a new subtype of migraine with aura and chronic migraine was recognized as a complication of migraine. The criteria for chronic migraine were later revised in 2006 resulting in a broader concept of this disorder.¹²⁴

A New Drug for Migraine – The Discovery of Sumatriptan (1988).— The prototypical triptan, sumatriptan, was discovered by a team headed by Patrick Humphrey and became available in Europe in 1991.^16,120,125 It represented the first experimentally based approach to novel treatment of acute migraine attacks.¹²⁶ Sumatriptan proved to be a highly effective (at least subcutaneously) and well-tolerated drug for the treatment of migraine attacks, and it was hailed as a medical breakthrough. The research was concentrated on the possible role of 5-hydroxytryptamine (serotonin) in migraine therapy as mentioned above in the section on methysergide. In 2 open studies,^127,128 intravenous serotonin (5-HT) was found effective in the treatment of migraine attacks, albeit with so many adverse events that its therapeutic use would be impracticable. The research team in England set out trying to find the 5-HT receptor type responsible for 5-HT's beneficial effect. Saxena had found that methysergide had a selectively constrictor effect in the dog carotid bed and suggested that this was an “atypical” 5-HT receptor.¹²⁹ As part of an investigation into the mode of action of antimigraine drugs, a study of the excitatory receptors for 5-HT was carried out in a range of isolated vascular preparations of dogs.⁶⁶ Serotonin was an agonist that resulted in contraction of all vessels whereas methysergide was an agonist only in the femoral vein.⁶⁶ It was hypothesized that this was an unknown 5-HT receptor in the dog femoral vein.

5-hydroxaminotryptamine (5-CONH₂T), a potent selective 5-HT agonist, had only a weak effect on rabbit isolated aorta, whereas 5-CONH₂T was a potent agonist in dog saphenous vein.¹³⁰ In this vein ketanserin, a 5-HT₂ antagonist, did not antagonize the effect of 5-CONH₂T. Thus, the receptor mediating contraction in the dog saphenous vein appeared to be “5-HT-like.”¹³⁰ Sumatriptan, which was synthesized in 1984,¹²⁶ appeared to have a selective effect on the dog saphenous vein and was accepted for clinical development on the basis of its high degree of selectivity for vascular “5-HT₁-like” receptors that mediate constriction.¹³⁰ These receptors are largely localized on large intracranial blood vessels from a variety of species including man^131-135 and sumatriptan causes contraction of these vessels via an action on the 5-HT-1B receptor.¹³⁶ The triptans, including sumatriptan, are relatively cranioselective when compared the effect on coronary arteries.^122,137 A possible central effect of the triptan is probably mediated by both 5-HT-1B and 5-HT-1D receptors and other 5-HT receptors.¹²²

The effect of subcutaneous sumatriptan 6 mg was proved in 2 large placebo-controlled, in-clinic RCTs. Headache relief rates of 70%¹³⁸ and 72%¹³⁹ after 1 hour were shown. Subcutaneous sumatriptan has a reasonable well-defined dose-response, with 1 mg being the minimum effective dose and 6 mg being the optimum dose with no gain by increasing to 8 mg.^138-141

Oral sumatriptan became available and has been the standard triptan, being compared with all new oral triptans and other nontriptans drugs.^122,142 Despite being not effective in some patients,¹²² sumatriptan and the other triptans (zolmitriptan, naratriptan, rizatriptan, almotriptan, eletriptan, and frovatriptan) have generally revolutionized the treatment of migraine. Recently, a fixed combination of sumatriptan 85 mg and naproxen 500 mg (32% pain-free [PF]) was superior to placebo (10% PF) and sumatriptan 85 mg (24% PF) in 2 very large RCTs (n = 1461 and n = 1495) and sustained PF for 24 hours was 24%, 8% and 15%, respectively.¹⁴³ From a clinical perspective it is the evaluation that despite highly statistically significant results in very large RCTs, the majority of the migraine patients are not treated satisfactorily with triptans, with 30-40% PF response at 2 hours with most triptans¹⁴²

Migraine and Calcitonin Gene-Related Peptide (CGRP) (1990).— In 1983, a novel neuropeptide, CGRP, was demonstrated in neural tissue¹⁴⁴ and its presence in perivascular nerves of cerebral arteries was demonstrated with immunocytochemistry and radioimmunoassay.¹⁴⁵ CGRP was found to be a potent vasodilator of cerebral vessels.^146,147 Stimulation of the human trigeminal ganglion in the treatment of trigeminal neuralgia resulted in flushing and the release of vasoactive peptides, substance P, and CGRP, in the external jugular vein (EJV).¹⁴⁸ In 1990 it was shown that CGRP, but not neuropeptide Y, vasoactive intestinal peptide, and substance P, was considerably increased in the EJV during migraine attacks both in migraine with and without aura.¹⁷ Three years later, the effect of trigeminal ganglion stimulation on CBF and jugular vein peptides in cats was studied before and after administration of sumatriptan and dihydroergotamine.¹⁴⁹ The increase of CBF and release of CGRP in EJV in cats was reduced by both drugs. Treatment of migraine patients with sumatriptan also led to a decrease of elevated CGRP in the EJV and relief of headache in most cases.¹⁴⁹ The finding of increased CGRP in the EJV led to the development of new migraine drugs based on CGRP receptor blockade.¹⁵⁰

However, in a Danish study (n = 17) with intra-patient comparison, in which blood samples from the EJV were taken in the patients' home, there was no tendency for an increase of CGRP during an attack of migraine without aura.¹⁵¹ In a later study, also with intra-patient comparisons, 8 migraine patients were investigated in the laboratory during, and outside, attacks of migraine without aura. No increase of CGRP in EJV was found.^152,153 Furthermore, in one nitroglycerin-induced migraine attack study, CGRP in EJV was not increased.¹⁵³ In contrast, saliva CGRP was increased during migraine attacks in patients responding to rizatriptan¹⁵⁴ whereas there was a nonattack-related increase in CGRP in saliva in migraine in another study.¹⁵⁵

The important role of CGRP in migraine pathophysiology is shown by 2 sets of facts. First, infusion of CGRP induced delayed migraine attacks in migraine patients.¹⁵⁶ Second, CGRP receptor antagonists were effective in the treatment of migraine attacks. This has been demonstrated in 4 placebo-controlled RCTs with the 2 CGRP receptor antagonists, the intravenous drug olcegepant and the oral drug telcagepant.^157-160 Because of the lack of cardiovascular adverse events of these CGRP antagonists,¹⁶¹ this is a major breakthrough in migraine therapy. Recent data, however, reported elevated transaminases when telcagepant was administered twice daily for 3 months for the prevention of migraine rather than acutely¹⁶² and the future of the drug is uncertain.

The Brainstem “Migraine Generator”– PET Studies in Migraine (1995).— In 1995, 7 patients with right-sided migraine without aura, were studied by PET within 6 hours after the onset of migraine symptoms, as well as outside attacks. During the attacks, increased blood flow was found in the cerebral hemispheres in the cingulate, auditory, and visual association cortices as well as in the brain stem (Fig. 9), slightly lateralized to the left.¹⁸ Only the brainstem activation persisted after subcutaneous sumatriptan had induced relief of symptoms. This was the first report of a strong brainstem activation in association with an acute, spontaneous attack in patients with migraine without aura. The authors stated that “it is tempting to consider the observed activation in the brainstem as the visualization of the postulated migraine centre in humans.”¹⁸ This activity in the brainstem has been termed the “migraine generator.”¹⁶³ In addition to spontaneous migraine attacks, changes in the brainstem have been studied with PET during nitroglycerin-induced attacks of migraine without aura.⁸¹ An activation lateralized to the side of headache was observed and this activation persisted after successful treatment with sumatriptan. In a study with PET in spontaneous migraine (n = 7) without aura attacks the brain stem activation was confirmed and in addition activation in the hypothalamus was observed.⁸⁰ It was suggested in one of the papers⁸¹ that lateralization of pain in migraine was due to lateralized brain dysfunction, and that the data reinforced the view of migraine as a brain dysfunction.

Figure 9.—. Regional cerebral blood flow (rCBF) PET measurements in 9 migraine patients.¹⁸ Data were calculated by subtracting rCBF during the headache-free phase from the migraine attack. Note the increase in rCBF in the brain stem. The activation in the singulum is unspecific. This brain stem activation persisted after subcutaneous sumatriptan 6 mg had induced complete relief from headache and phono- and photophobia. It has been suggested that this brain stem activation could be a “migraine generator.”¹⁵⁷

Migraine as a Channelopathy? Research From the Genetic Perspective (1996).— A major scientific breakthrough in migraine research was the introduction of genetic studies. This included both traditional clinical genetic methodologies, basic genetic research, and pharmacogenomics. The heredity of the 2 forms of migraine was most likely different. Thus, in a population-based study the first degree relatives of probands of migraine without aura had 1.9 times the risk of migraine without aura (compared with the general population) and 1.4 times the risk of migraine with aura.¹⁶⁴ The first degree relatives of probands of migraine with aura had nearly 4 times the risk of migraine with aura and no increased risk of migraine without aura.¹⁶⁴ The regional cerebral blood changes were different in the 2 forms of migraine,^12,75,76ut supra. The phenotype is most likely different for the 2 forms of migraine. However, some investigators suggest that there is a migraine continuum.^165,166

In 1993 the linkage of the monogenic FHM to chromosome 19p13 was reported by the Paris group.¹⁶⁷ This was soon followed by proof of genetic heterogeneity because approximately only 50% of FHM families appeared to be linked to this locus.^168,169 In a subsequent study, including 4 families (3 with migraine with aura and 1 with migraine without aura) with multiple cases, the reported locus of FHM was excluded.¹⁷⁰ In contrast, another study from 1995 suggested that the FHM locus on 19p13 was involved in the common forms of migraine with and without aura.¹⁷¹ In 1996 the protein kinase substrate 80 K-H gene was excluded as a candidate gene for FHM by the Leiden group.¹⁷² Then later in 1996 a seminal paper in Cell from the Leiden group reported the first missense mutation in the voltage-gated P/Q Ca2⁺ channel CACNA1A gene in FHM¹⁹ (see Fig. 10). They examined 16 patients with FHM and 50 randomly collected controls with Exon trap experiments, cDNA sequence, and Northern blot analysis. Genomic DNA was used as a template to generate polymerase chain reaction products for single-strand conformational polymorphism analysis and denaturing high-performance liquid chromatography. The authors concluded that “our findings implicate the P/Q-type channel α1-subunit gene on chromosome 19p13.1 (CACNL1A4) in the pathogenesis both episodic ataxia type 2 and FHM, and most likely also of the more common forms of migraine.”¹⁹ A variable expression of mutations in the P/Q-type calcium channel was observed.¹⁷³ Since then, 2 other mutations for FHM, the ATP1A2 gene located on chromosome 1q23 (FHM2)^173,174 and the SCN1A gene located on chromosome 2q24 (FHM3)¹⁷⁵ have been identified.¹⁷⁶

Figure 10.—. Membrane topology of α_1A subunit of the P/Q-type Ca²⁺ channel, CACNA1A. The location and amino acid substitutions are indicated for mutations that cause familial hemiplegic migraine (FHM), episodic ataxia type 2 (EA-2), tottering mouse (tg), and spinocerebellar ataxia type 6 (SCA6).¹⁹

Since 1996, several studies have shown linkage to other loci or genes in migraine with and without aura.^177-183 In contrast, the D₂ receptor Ncol allele did not have an allelic association of migraine with aura.¹⁸⁴

In one large genetic study in patients (n = 827) with “typical migraine” and controls (n = 765) single-nucleotide polymorphism (SNP) alleles in the insulin receptor gene were associated with migraine.¹⁸⁵ A replication study (949 patients and 648 controls) in migraine with aura showed, however, only a nonsignificant trend for an SNP in the insulin gene and migraine (P = .1).¹⁸⁶

In a recent comprehensive and large-scale study including 2800 migraine with aura patients from various countries, it was tested whether common variants in ion transport genes could be involved in a common type of migraine.^187,188 More than 5000 SNPs in 155 in transport genes (including the 3 FHM genes) were studied, but no significant associations were found.^187,188 From this study it seems that common variants in ion transport genes do not play a major role in susceptibility for common types of migraine. Thus, the large majority of genetic associations with candidate migraine genes have not been convincingly replicated, nor have any gene mutations from these loci been published (see recent review¹⁸⁸).

One reason for the lack of establishing an association between the FHM genes and common types of migraine could be different phenotypes. Thus, neither FHM1 or FHM2 showed hypersensitivity to nitric oxide^189,190 as observed in migraine with and without aura.^117,191 The major contribution of the identification of the gene for FHM was that it had inspired many investigations of the genetics of the more common types of migraine up to the present.

Meningeal Sensitization, Central Sensitization, and Allodynia in Migraine (1996).— In an influential animal study, it was shown that brief chemical irritation by inflammatory soup (IS) (histamine, serotonin, bradykinin, prostaglandinE₂) on the dura in rats resulted in sensitization of the primary meningeal afferent nerves.¹⁹² In another rat study, brief exposure of the dura to chemical agents associated with inflammation, including IS, caused central dura-sensitive neurons to become more sensitive to mechanical stimulation of the dura and to mechanical and thermal stimulation of the head's skin.¹⁹³ This indicated sensitization of central trigeminal neurons that receive convergent input from the dura and the skin.¹⁹³ This central sensitization could most likely explain allodynia observed in migraine.¹⁹³ In a recent large population-based study in 11,388 migraine patients, the prevalence of central sensitization was 63% and severe cutaneous allodynia (CA) occurred in 20% of migraineurs.¹⁹⁴ A logical following step was investigating CA prospectively in migraine patients.²⁰ Pain threshold for heat, cold, and mechanical stimuli were determined outside and during migraine attacks. The regions examined were the left and right periorbital skin, as well as the ventral skin of the left and right forearm. It was found that during migraine 33 of 42 patients experienced various patterns of CA. Twenty-eight patients had allodynia beyond (1-3 sites) the ipsilateral side of the head suggesting that this was due to sensitization of at least third-order neurons that receive convergent input from the dura, periorbital skin, and skin areas of various body sites.²⁰ In one subject, who was observed during an untreated migraine attack, there was clinical evidence, development of allodynia, for the sequential recruitement of second-order and third-order neurons in migraine.¹⁹⁵

The influence of presence of allodynia on the effect of triptans was investigated by treating early (<1 hour) or later (4 hours) from the onset of attack in an open crossover study.¹⁹⁶ Within 2 hours of triptan treatment, patients were rendered PF after 2 hours in 5 of 34 (15%) of allodynic attacks vs 25 of 27 (93%) of nonallodynic attacks.¹⁹⁶ The authors concluded that the probability of consistent PF outcome is increased drastically if triptan therapy is vigilantly timed to precede any signs of allodynia.¹⁹⁶ Similarly, in animals early treatment with sumatriptan prevented initiation of central sensitization triggered by chemical stimulation of meningeal nociceptors whereas late treatment was insufficient to counteract an already established central sensitization.¹⁹⁷ In one animal study it was shown that triptans disrupted communication between peripheral and central trigeminovascular neurons.¹⁹⁸

In a group of 28 migraine patients with allodynia treated 4 hours after onset, subcutaneous sumatriptan was ineffective in 14 patients. These patients were subsequently treated i.v. with the COX1/COX2 inhibitor ketorolac as were 14 other allodynic patients.¹⁹⁹ A PF state was observed in 71% and 64%, respectively.¹⁹⁹ In animal studies COX1/COX2 inhibitors, ketorolac, indomethacin, and naproxen, can exert inhibition of the central trigeminovascular neurons and can suppress central sensitization in rats.²⁰⁰ In contrast, subcutaneous sumatriptan was equally effective when given early (62% PF after 2 hours) or late (55% PF after 2 hours) during migraine attacks in one open study (n = 20).²⁰¹ In an unpublished RCT (n = 90) both early and late (4 hours) treatment of migraine attacks with subcutaneous sumatriptan 6 mg resulted in headache relief in about 80% in both groups.²⁰¹ In addition, subcutaneous naratriptan 10 mg in an RCT resulted in 88% of patients (n = 34) being PF after 2 hours even if more that 50% were treated after 4 hours.^202,203 These 2 studies indicate that parenteral triptans are equally effective when used early or late in the treatment of migraine attacks.

Another way to circumvent the problem of allodynia is to treat early which makes common sense. In one prospective, placebo-controlled RCT with almotriptan 12.5 mg it was demonstrated that early (<1 hour) and mild headache treatment (53% PF) was superior to treatment of moderate or severe headache (38% PF).²⁰⁴

Citations of Highlighted Papers.— There was comparable, and reasonable, agreement between ISI Web of Knowledge and Google Scholar (see Table 2). The 2 papers with most citations, are Leão's paper (1129) on CSD¹⁰ and the Leiden Group's paper (1215) on the gene for FHM1.¹⁹

SUMMARY AND CONCLUSION

In this review we tried to highlight the major clinical and scientific observations of migraine from 1910 to 2010. This was undertaken from the present perspective and therefore observations that may have seemed important in the past may have been omitted, although we believe we have touched upon the main areas that have occupied migraine investigators. After the discovery of ergotamine in the late 19th century, its isolation in the early 20th century led to rather primitive trials (from today's perspective) as well as research on the pathogenesis of migraine. American research of therapeutics in the 1930s seemed more thorough than the early European endeavors. Striking differences in the study of ergotamine in the 1920s and 1930s vs sumatriptan in the 1980s reflect the evolution of classifications necessary for better definitions and uniform patient groups, on the one hand, and the improved knowledge in clinical epidemiology on the other.

Although there had been some confusion on whether migraine was a cerebral vasoconstrictor or vasodilator disorder early in the century, the work by Wolff et al in the 1930s further established the vasodilator model, in which the vasodilatory action could be counteracted by ergotamine. The vascular vs the neurogenic theory had been debated previously for more than 100 years²⁹ and is still debated up to this day.^43,205

Modern neurosurgery in the early 20th century resulted in new questions with respect to pain eliciting structures within the skull, but also made it possible to perform such research on migraine patients. These studies were not only interesting with respects to the results, but also from the perspective of ethics of the period, subjecting volunteer patients to painful experiences during cranial surgery. An important aspect of migraine that became better understood around the middle of the 20th century was the visual aura that had occupied so many 19th century scientists, who had observed the phenomenon themselves. However, only after the advancement of neurophysiological knowledge, Karl Lashley was able to advance knowledge about visual auras and to determine the speed of progression of an inhibitory or excitatory process over the occipital cortex. It was not immediately linked to the phenomenon of CSD that was discovered at about the same period, the investigators being unaware of Lashley's description. Only after 1980 was the possible relationship between CSD and aura, with spreading oligemia, considered likely by the application of rCBF in human as well as animal studies. The vascular hypothesis received several new impulses from the field of neurochemistry leading to new prophylactic and ultimately acute treatments. Finally, the discoveries in the field of genetics brought migraine (at least hemiplegic migraine) within a new system of diseases, which provided a new perspective.

Pathophysiological ideas on migraine evolved within a limited number of paradigms, notably the vascular, neurogenic, neurotransmitter, and genetic/molecular biological paradigm. The application of various new technologies played an important role within these paradigms, in particular neurosurgical techniques, EEG, neurochemistry, methods to measure rCBF, PET imaging, clinical epidemiological, genetic, and molecular biological methods, the latter putting migraine (at least hemiplegic migraine) in a whole new system of diseases.

STATEMENT OF AUTHORSHIP

Category 1

• (a) 
  Conception and Design
  Peer C. Tfelt-Hansen; Peter J. Koehler
• (b) 
  Acquisition of Data
  Peer C. Tfelt-Hansen; Peter J. Koehler
• (c) 
  Analysis and Interpretation of Data
  Peer C. Tfelt-Hansen; Peter J. Koehler

Category 2

• (a) 
  Drafting the Manuscript
  Peer C. Tfelt-Hansen
• (b) 
  Revising It for Intellectual Content
  Peer C. Tfelt-Hansen; Peter J. Koehler

Category 3

• (a) 
  Final Approval of the Completed Manuscript
  Peer C. Tfelt-Hansen; Peter J. Koehler