Abstract
Sjaastad and Dale (1) reported a new treatable headache entity in 1974, and in 1976 they named the entity chronic paroxysmal hemicrania (2). From the beginning, the discoverers insisted on the absolute effectiveness of indomethacin in PH, and were effective in ‘splitting’ the disorder from cluster headache. Professor Sjaastad described the complete indomethacin response in CPH as a
‘dramatic and flabbergasting event, to the extent that it seems unbelievable to both the patient and the physician (3).’
The first case seen by Professor Sjaastad was described in detail in his book Cluster Headache Syndrome and will be quoted extensively because it clearly demonstrates the dramatic therapeutic effect of indomethacin (3).
‘The first case of chronic paroxysmal hemicrania (CPH), a female aged 44 years with a 9-year history of headache, was brought to our cognizance in 1961, with a diagnosis of ‘typical cluster headache’. However, the patient was a female (which, of course, may be consistent with a diagnosis of ordinary cluster headache, although females are admittedly rarely affected). There was another trait that, upon scrutiny, did not seem to be quite typical: there was a multitude of attacks per 24 h, i.e. up to 24 or more per day. Another remarkable feature was the intractability of the headache. Over the course of the following years, every feasible drug was tried on her. Prior to each admission, she had discontinued her usual drug (mainly acetylsalicylic acid) which kept the attacks at a reasonable level. Each drug trial ended either with an absolutely negative response, or – more usually – with an adverse reaction. The latter response pattern was so typical that for almost every new drug that was tried she was brought from a stage of moderate or weak attacks to a stage of incapacitating and excruciatingly severe attacks. But she would not give up the hope that there was a solution to her problem. And she was the one who motivated us to try new approaches, not the other way around. To make matters even worse, she was diagnosed as being an hysterical person … despite the unilaterality of the pain, lacrimation, rhinorrhoea, etc.
The patient felt that everything centred on the eye. We had arrived at a point where we were running out of arguments when trying to reject the patient's contention that an enucleation of the right (symptomatic side) eye would perhaps solve her pain problem.
Fortunately, the following developments took place simultaneously. The patient had from the start claimed that salicylates seemed to give her some relief, not to the extent that the pain disappeared, but at times when they were not maximal the paroxysms could be modified by salicylates. Perhaps the paroxysms became even more rare during such medication. We thought that a lot of drugs would be more promising than salicylates as potential therapeutic agents. The anti-inflammatory agents were therefore put rather far down on our list of potentially effective drugs worthy of trial. When in late 1972 we finally arrived at indomethacin on our list, the response was no less than miraculous (3).
A remitting form of the disease was subsequently recognized and termed episodic paroxysmal hemicrania (4). The International Headache Society gave diagnostic criteria for CPH in 1988 (5), and in 2004 the IHS classified CPH and EPH as subtypes of paroxysmal hemicrania (6).
PH was initially felt to be a disease of women, but over time the female:male ratio has decreased (7). The typical clinical features of PH became established as the number of reported patients increased. PH attacks are classically shorter in duration and higher in frequency than cluster attacks. The majority of PH patients have a chronic pattern, unlike cluster headache. These differences are not absolute, however, and clinicians have noted a considerable overlap in the clinical characteristics of PH and cluster headache (8, 9).
Zidverc-Tajikovic et al. (10) compared the clinical features of 54 cluster headache patients with 8 PH patients. Differences were found in maximal pain localization (extra-ocular in PH), mean attack duration (shorter in PH), mean attack frequency (higher in PH), and clinical course (more often chronic in PH). There was no clinical characteristic that exclusively belonged to one headache type. The response to verapamil did not differ significantly between the two groups. Verapamil-responsive PH has been reported previously (11). All PH patients responded to indomethacin in daily dosages of 75–150 mg. However, it is not clear that the indomethacin nonresponders were titrated up to 150 mg daily. Thus, some of the cluster headache patients may have had PH if doses of less than 150 mg daily were tried or if an Indotest was not performed (12). The authors conclude that the therapeutic response to indomethacin is the most reliable differential diagnostic criterion for cluster headache and PH.
Another possibility is that there is a subtype of cluster headache responsive to indomethacin. However, the simplest explanation is that some PH patients (as defined by a complete indomethacin response) have a phenotype more commonly associated with cluster headache. Indomethacin has been shown ineffective in the treatment of cluster headache in one uncontrolled study (13). In clinical practice, if a patient with a phenotype suggestive of PH fails an adequate indomethacin trial, it is more rewarding to next try medications typically efficacious in cluster headache than to try other medications reported to be helpful in PH.
I and other authors previously argued that indomethacin response should not be required as a diagnostic criterion for PH (14, 15), but over time I have come to realize this necessity. Given the phenotypic overlap between cluster headache and PH, requiring an indomethacin response is the only way to identify a group that can be rigorously studied to determine the aetiology of the disorder. This is analogous to dopa-responsive dystonia, where treatment response lead to the splitting of a disease from a larger phenotypic group (16). The mechanism by which indomethacin works in PH is unclear. It appears to be independent of its effect on prostaglandin synthesis, since other NSAIDs are not usually efficacious (17). Unlike other NSAIDs, indomethacin can decrease intracranial pressure (17), but that seems unlikely to be important in PH given that intracranial pressure was not elevated when measured during an attack of PH (18). This latter study was reported in 1970 as being done on a patient with cluster headache (18), but this patient was eventually diagnosed with PH after having a complete indomethacin response (3). One patient with CPH and a CSF opening pressure of 30 cm H2O has been reported, but that case was unusual in that the patient had side-alternating attacks that required indomethacin doses of 275 mg per day to control, and no comment was made about the presence or absence of papilledema (19). Indomethacin inhibits the production of nitric oxide (NO), which is significant given that glycerol trinitrate (a NO donor) can trigger neurovascular headaches like cluster headache and migraine (17). NO likely activates CGRP release from trigeminal fibers (20). NO is colocalized with VIP within neurones and nerve fibers of the cranial parasympathetic system (17). The elevated CGRP and VIP levels seen during attacks of PH return to normal after treatment with indomethacin (21). Thus indomethacin may exert its effect on PH by antagonizing one or more steps in the NO pathway, since PH is a disorder characterized by activation of the cranial parasympathetic system (17).
In 1992, Professor Sjaastad wrote that ‘indomethacin in CPH may ‘hit the nail on the head’ (3).’ This indomethacin ‘hammer’ would appear to be the best tool for the job of differentiating paroxysmal hemicrania from cluster headache.