Cluster headache
CH is the most frequent syndrome. The pain is located mainly around the orbital and temporal regions, though any part of the head can be affected. The headache usually lasts 45–90 min but can range between 15 min and 3 h. Typically, this syndrome is characterized by a striking circannual and circadian pattern. There is a clear male preponderance.
Usually headache attacks cluster in time, lasting for 7 days to a year separated by remission periods lasting for months or years (80–90% of patients). Sometimes attacks recur for more than 1 year without remission periods, then becoming chronic (10–20% of patients) [
21,
22].
Early neuroimaging studies in cluster headache evaluated cerebral blood flow, using mainly SPECT. This semi-quantitative method has not provided univocal results, since some studies reported an increase, some a decrease and others no differences in cortical blood flow, possibly because of methodological dissimilarities [
23‐
27]. Di Piero et al. [
28], aiming to investigate brain response to pain in 7 cluster headache patients out of the bout compared to 12 healthy controls, recorded Xe-133 SPECT during experimentally induced pain by means of a cold water pressor test. They demonstrated less cerebral blood flow modifications in contralateral primary sensorimotor and thalamic regions compared to healthy subjects only when the CPT was performed on the CH side. This led the authors to suggest the possible involvement of central tonic pain mechanisms in the pathogenesis of cluster headache. Hsieh and coworkers performed [
15O] butanol PET in 7 patients affected by episodic CH (4 in and 3 out of the bout) during nitroglycerine-induced pain. PET scan showed a significantly increased rCBF in the right caudal and rostrocaudal anterior cingulate cortex (ACC), temporopolar region, supplementary motor area, bilaterally in the primary motor and premotor areas, opercular region, insula/putamen, and lateral inferior frontal cortex. Moreover, a reduction in rCBF bilaterally in the posterior–parietal cortex, occipito-temporal region and prefrontal cortex was observed [
29]. May and colleagues, scanning nine chronic CH patients with H
2
15
O PET during nitroglycerine-induced attacks, were the first to clearly demonstrate inferior hypothalamic gray matter activation ipsilateral to the headache side. Moreover, they observed an increased rCBF in the contralateral ventroposterior thalamus, the anterior cingulate cortex, and in the insulae bilaterally as well [
30]. Later, other authors confirmed these data in a spontaneous headache attack of a chronic CH patient during an ongoing H
2
15
O PET study [
31]. The CH attack-induced activation also increases in the medial thalamus and contralateral perigenual ACC [
31].
The exact role of the intracranial blood vessels in the mechanism of cluster headache was investigated by May et al. [
31], who performed a H
2
15
O PET study in a group of 17 episodic CH patients (9 in the active period and 8 out of their bout) and a MR angiography in a spontaneous CH and capsaicin-induced pain in 4 healthy volunteers. H
2
15
O PET study showed significant activation during spontaneous or nitroglycerine-induced headaches bilaterally in the ACC, ipsilateral posterior thalamus, ipsilateral basal ganglia, ipsilateral inferior posterior hypothalamus, both frontal lobes, bilaterally in the insulae, and in the contralateral inferior frontal cortex. In addition, they found activation in large intracranial vessels on the PET scan which corresponds to a significantly increased blood flow in the internal carotid artery ipsilateral to the headache side, both in CH patients and in experimentally induced pain [
32]. The latter findings further support the neuronal nature of the dysfunction in CH, confirming that intracranial vascular changes are not specific to headache but represent a generic epiphenomenon of the pain.
Sprenger and colleagues [
4] measured cerebral glucose metabolism by means of FDG-PET in 11 episodic CH patients during and outside the bout and compared these patients with a group of 11 healthy subjects. With respect to those outside the bout, patients scanned during the bout presented increased metabolism in the perigenual ACC, posterior cingulate cortex, the orbitofrontal cortex including the nucleus accumbens, ventrolateral prefrontal cortex, dorsolateral prefrontal cortex (DLPFC) and temporal cortex, and increased metabolism in cerebellopontine area. Moreover, CH patients (in and out of the bout) compared to healthy subjects revealed hypometabolism in the perigenual ACC, prefrontal and orbitofrontal cortex [
4]. Interestingly, the perigenual ACC was found to be activated in chronic CH patients, unresponsive to pharmacological therapy, who were treated successfully with occipital nerve stimulation [
33].
Further evidence for hypothalamic dysfunction in CH arises from spectroscopic studies. A study with proton MR spectroscopy (
1H-MRS) of 26 patients with CH (18 episodic, 10 in and 8 outside the bout, and 8 chronic) demonstrated that the NAA, a marker of neuronal integrity, is reduced in the hypothalamus of three subgroups when compared to 12 healthy subjects [
34]. These data have been confirmed with the same methodology in a group of 47 episodic CH patients by Wang and colleagues [
35], who found in addition, a reduction in the Cho/Cr metabolite ratio, both during and out of the bout. This suggests that the hypothalamus in cluster headache might be characterized not only by a neuronal dysfunction but even by changes in the membrane lipids.
With the voxel-based morphometric (VBM) analysis, May and coworkers [
36] studied 25 episodic patients with CH and reported an increase in bilateral hypothalamic gray matter volume, with similar results in patients examined during and outside the bout [
36]. Another larger VBM study investigating 75 CH patients (22 episodic inside bout, 35 outside bout and 18 chronic CH) was unable to reveal changes in hypothalamic area neither overall nor in subgroups [
37].
VBM-MRI findings taken together with those provided by
1HMRS may indicate that the hypothalamus of patients with CH has an increased neuronal density with reduced NAA, suggesting the presence of either immature or dysfunctional neurons [
34,
38]. Moreover, these morphometric and functional changes seem to be a permanent disease-related dysfunction since they are not due to the CH history or the cluster phase [
34,
36].
Using PET with the opioidergic ligand [11C]diprenorphine in 7 CH patients (6 episodic and 1 chronic) who are in bout but out of an acute attack, Sprenger and colleagues demonstrated a decreased tracer binding in the pineal gland, but not in any other brain structure commonly claimed to be involved in cluster headache pathophysiology [
7,
36]. Furthermore, the authors found an inverse relationship between the duration of cluster headaches and opioid receptor binding in the ipsilateral hypothalamus and bilateral cingulate cortices. The latter observation suggests that descending opioidergic mechanisms in the pineal gland and hypothalamus might be involved in the generation of cluster headache attacks.
The study of Morelli et al. [
39] is the only one performing fMRI in order to investigate the pattern of cerebral activation during an attack of CH in four episodic patients, showing activation of diencephalic regions, mainly the hypothalamus. Additionally, they documented trends of activation in cerebral areas involved in pain processing (prefrontal cortex, anterior cingulate cortex, contralateral thalamus, ipsilateral basal ganglia and the insula and the cerebellar hemispheres bilaterally) [
39].
Paroxysmal hemicranias
PH is a relatively rare syndrome and the clinical features are highly characteristic [
40‐
42]. Patients typically have unilateral, brief, severe attacks of pain, localized over the first division of the trigeminal nerve, associated with cranial autonomic features that recur several times per day. The duration of the pain is between 2 and 30 min with a frequency of between 1 and 40 attacks per day. It is more common in women and in chronic form. PH responds in a dramatic and absolute fashion to indomethacin, hence the importance of distinguishing it from CH and SUNCT, which are not responsive to indomethacin. Activation of the hypothalamic gray matter during attacks was also observed in paroxysmal hemicranias. Matharu and coworkers [
43] were the first to perform H
2
15
O PET in seven patients affected by PH, scanned during acute attack-off indomethacin or pain-free-off indomethacin and pain-free due to indomethacin administration. The regions significantly activated during headache-off indomethacin versus the pain-free phase were the contralateral posterior hypothalamus, contralateral ventral midbrain, ipsilateral lentiform nucleus, anterior and posterior cingulate cortices, bilateral insulae, bilateral frontal cortices, contralateral temporal cortex, contralateral postcentral gyrus, precuneus, and contralateral cerebellum [
43]. Interestingly, indomethacin administration turned off the persistent activation observed during acute attack-off indomethacin. This study supports the view of paroxysmal hemicrania as a central nervous system disorder and demonstrates for the first time the metabolic correlation with indomethacin efficacy.
Short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing
Among the other TACs, SUNCT is a very rare primary headache syndrome, with a male prevalence, characterized by strictly unilateral, severe, neuralgic attacks centered on the ophthalmic trigeminal distribution in association with both conjunctival injection and lacrimation [
21,
44]. The duration of each attack is between 5 and 240 s with a frequency ranging between 3 and 200 per day. Although there are marked differences in the clinical pictures of the trigeminal autonomic syndromes, such as the frequency and duration of attacks and the different approaches to treatment, many of the basic features of SUNCT, such as episodicity, autonomic symptoms, and unilaterality, are shared by other headache types, such as cluster headache and chronic paroxysmal hemicrania suggesting a pathophysiological similarity to these syndromes. By studying 6 consecutive spontaneous attacks of SUNCT with fMRI in a 71-year-old woman, May et al. [
45] observed activation during the same scanning session in the ipsilateral inferior posterior hypothalamic gray matter during the attacks, compared with the pain-free state. The same ipsilateral area activation was observed also in a 68-year-old man with an atypical case of TAC resembling SUNCT [
46]. Additionally, a clear trend of cingulate cortex, insula, temporal and frontal cortices activation was observed [
46]. It has to be noted that activation of the posterior hypothalamus was even bilaterally in the two patients affected by SUNCT in the study by Cohen and colleagues [
47] and increases parametrically with increasing levels of pain.
Bilateral posterior hypothalamus activation was also observed in the fMRI study of Sprenger et al. [
48] where a 49-year-old male patient was scanned during attacks of SUNCT and pain-free state. Also in this case, multiple activations were observed in brain regions involved in the processing of pain [
48].