Abstract
Chronic paroxysmal hemicrania (CPH) is one of a group of primary headaches disorders known as the trigeminal autonomic cephalgias (TACs) (1), which also includes the more common entity of cluster headache. These headaches share a number of clinical characteristics, such as their unilateral first and second division trigeminal pain and ipsilateral cranial autonomic activation. They also probably share pathophysiological similarities as demonstrated by recent functional imaging studies. Clinical differentiation between the two entities may be difficult because of the phenotypic overlap between them before a trial of medication (2, 3). This differential diagnosis hinges on important and intriguing differences between them in their response to medication. The pain and autonomic features of CPH respond uniquely and absolutely to indomethacin (4) but not triptans, whereas most cluster headache sufferers have a striking benefit with subcutaneous sumatriptan but no response to indomethacin.
It is well known that cluster headache is readily triggered by nitric oxide (NO) donors, e.g. glyceryl trinitrate (GTN), during active headache periods (5), but not during the asymptomatic interval (5). GTN triggering of CPH has not previously been reported. This report concerns a 78-year-old lady whose attacks of CPH could be triggered by GTN. Indomethacin abolished not only her spontaneous CPH attacks but also those triggered by GTN.
Case report
A 78-year-old woman gave a 2-year history of piercing right-sided headaches. Very occasionally they affected the left side of the head. The pain lasted from 5 to 9 min and occurred two to seven times a day. She graded the pain as 8/10 in severity. There was no associated nausea, photophobia or phonophobia. During each attack her eye watered but no other autonomic symptoms had been noted. She tended to sit still during the attack. Paracetamol did not relieve the pain. She had a past medical history of hypertension and angina for which she used GTN sublingual spray. The GTN spray triggered her typical attacks of pain within a few minutes. She was on no other regular nitrates. She had a past history of bilateral, milder headaches 20 years previously. She drunk minimal alcohol and did not know whether alcohol triggered attacks of head pain. An unenhanced magnetic resonance imaging brain scan was normal apart from selective frontal and temporal lobe atrophy, and areas of altered white matter signal in both hemispheres, mid-brain and pons, consistent with small vessel ischaemia. For these right-sided headaches she was treated with indomethacin 25 mg three times a day which completely relieved the attacks of pain, confirming the clinical diagnosis of CPH. She preferred to take just 25 mg once a day with 90% relief of the pain. Indomethacin not only prevented the spontaneous attacks, but also completely prevented the GTN-induced attacks.
Discussion
This is the first case report of a NO donor triggering CPH. This extends the clinical characterization of CPH. It adds to the list of clinical similarities between the TACs cluster headache and CPH and may give some insights into the mechanism of action of indomethacin in CPH.
Intriguing similarities have been found between CPH and cluster headache on functional imaging that may relate to a shared pathophysiology: a region of the posterior hypothalamus is activated ipsilaterally in spontaneous and experimentally induced cluster headache triggered by GTN, in addition to other cortical and brainstem sites known to be involved in pain processing (6, 7). A very similar region is activated, albeit contralaterally, in spontaneous CPH attacks (8).
How does NO trigger headache in cluster headache and in this case of CPH? It is known that GTN induces an immediate (vasodilatory) headache in most people and also a delayed (5–6 h post dose) headache in migraineurs: a time course more in keeping with second messenger systems than a direct vasodilatory effect. The headache induced in cluster by NO appears after a latency of 30–50 min. In the case reported, headache occurred within minutes of GTN, suggesting a direct vasodilatory effect.
It is thought that NO mediates release of calcitonin gene-related peptide (CGRP), from trigeminal fibres (9), which may, amongst other actions, induce arteriole dilation (10) and may be one mechanism by which NO induces headache. However, there are no experimental data relating NO to CPH and no good evidence as yet for vasodilation in CPH to substantiate this. One study (11) has described a case of CPH with elevated levels of CGRP in the cranial circulation during attacks. Similar neuropeptide changes are seen in cluster headache. This suggests that a degree of vasodilation mediated by CGRP may be present in CPH.
It is also not known how indomethacin alleviates the paroxysmal hemicrania syndrome, but not cluster headache. Indomethacin inhibits the production of NO by endothelial and inducible nitric oxide synthase (NOS) (12, 13), but not neuronal NOS (12). However, it is unlikely that this NOS inhibition is the mechanism of action of indomethacin in CPH, as this case report demonstrates that indomethacin can inhibit triggering of pain by a NO donor, i.e. an effect ‘downstream’ of NOS. In the study in which elevated levels of CGRP were found in the cranial circulation during CPH (11), successful treatment of the headache with indomethacin normalized the levels of CGRP. Furthermore, a previous experimental study has shown that both sumatriptan and indomethacin inhibited the NO-induced dilation of the dural meningeal vessels, with a likely action via trigeminal neurons (14). However, more work is necessary in order to understand the importance of CGRP-related mechanisms and the mode of action of indomethacin in the paroxysmal hemicrania syndrome. The absolute and unique action of indomethacin on CPH implies a very specific mode of action pertinent just to the pathophysiology of CPH.
Although functional imaging studies have shown striking similarities between cluster and CPH, there are clearly a number of important, as yet undiscovered, pathophysiological differences and the differing response to indomethacin may provide a vital clue.
This is the first case report of CPH attacks induced by GTN and underlines that CPH is part of the TACs, in spite of the very peculiar characteristic that indomethacin aborts the attacks and triptans are ineffective. If the reproducibility of GTN as a trigger of paroxysmal hemicrania headache is verified in a large group of patients, it may be possible to study experimental models in order to provide more insights into the neuroanatomical correlates of this rare primary disorder.