Abstract
Bingel and Weiller (1) report a case of short-lasting, frequent, bilateral, indomethacin-responsive headache without cranial autonomic features in this issue of Cephalalgia. They consider this to be a bilateral form of paroxysmal hemicrania (PH). We note two other case reports of short-lasting, frequent, bilateral, indomethacin-responsive headaches without cranial autonomic features that have been presented as bilateral PH (2, 3). These three case reports pose an interesting question with regard to their classification: do they represent a bilateral form of PH or are they cases of another novel indomethacin-responsive primary headache syndrome?
PH usually begins in adulthood at the mean age of 34 years and a range of 6–81 years (4). The condition predominates in females by a sex ratio of 1.6–2.4 : 1 (4–6). The clinical phenotype of PH is highly characteristic (4–8). The headache is strictly unilateral and without side shift in the majority of patients. The maximum pain is most often centred on the ocular, temporal, maxillary and frontal regions. The pain is described as a sharp, throbbing, stabbing, aching or boring sensation. The pain is typically excruciating in severity. The headache usually lasts 2–30 min, though can go on for up to 2 h. Attacks of PH invariably occur in association with ipsilateral cranial autonomic features. Lacrimation, conjunctival injection, nasal congestion, or rhinorrhoea frequently accompany the headache; eyelid oedema, ptosis, miosis and facial sweating are less frequently reported. Interestingly, there is a case description of a patient with otherwise typical PH, including a good indomethacin response, with no autonomic features (9). In addition, there are two further case reports of patients who had PH without cranial autonomic features though it is not clear whether these patients responded to indomethacin (6). Photophobia and phonophobia frequently accompany attacks, although nausea and vomiting are less common (6). There is one case report of a typical migrainous aura occurring in association with PH attacks (10). During episodes of pain, approximately half the sufferers prefer to sit or lie still while the other half assume the pacing activity usually seen with cluster headaches (5, 6). The attacks occur at a high frequency, ranging from two to 40 daily. The attacks occur regularly throughout the 24-h period without circadian periodicity or a preponderance of nocturnal attacks. However, nocturnal attacks associated with rapid eye movement (REM) phase of sleep have been described (11). While the majority of attacks are spontaneous, approximately 10% of attacks may be precipitated mechanically, either by bending or by rotating the head. Attacks may also be provoked by external pressure against the transverse processes of C4–5, C2 root, or the greater occipital nerve. Alcohol ingestion triggers headaches in only 7% of patients. PH responds in a dramatic and absolute fashion to indomethacin.
In the three reported cases of bilateral PH, the mean age of onset was 38 years with a range of 16–65 years. Two cases occurred in females and one in a male. The headache was bilateral with equal prominence on both sides in all three cases. The maximum pain was centred on the forehead, temples and the occiput. The pain was described as throbbing, piercing or crackling. It was intense or very severe. The duration of the headaches ranged from several seconds to up to 30 min. The frequency of the attacks ranged from two to 24 daily. Circadian periodicity was reported with some attacks in one patient and a predictable nocturnal attack was reported with another patient. Unlike in PH, the attacks were not associated with cranial autonomic features in any of the three patients. Photophobia and phonophobia accompanied the attacks in one patient; nausea, vomiting, osmophobia or aura symptoms were not reported in any patient. During episodes of pain, two of the patients were restless while one preferred to sit still, though movement did not worsen the pain. None of the patients identified any obvious triggers or relieving factors. The attacks were completely suppressed by indomethacin at a dose of 25–75 mg daily.
There are several similarities in the phenotypes of PH and the cases reported as bilateral PH. These include the mean age of onset, female preponderance, the site and character of pain, the short duration and high frequency of the attacks, and the absolute response to indomethacin. However, the three cases discussed here differ from PH in lacking unilaterality and the associated cranial autonomic features. There are rare reports of unilateral PH without cranial autonomic features. However, it is remarkable that there is a lack of cranial autonomic features in all the bilateral cases that are fully reported in the literature, thereby suggesting that this is highly unlikely to have occurred just by chance. In addition, there was a paucity of migrainous symptoms in the bilateral cases, in contradistinction to the reported phenotype of PH. Are these differences sufficient to consider classifying the bilateral headache phenotype as a novel indomethacin-responsive primary headache syndrome rather than a variant of PH? The answer to this question depends on the basis of classification of these disorders.
The classification of disorders can be based on the underlying pathophysiology, clinical phenotype or a specific treatment response. The classification of primary headache disorders would ideally be based on the underlying pathophysiology. However, the pathophysiological basis of most primary headache syndromes is poorly understood and therefore has limited utility as yet. As our understanding of the underlying basis of these syndromes improves, however, it is likely that this basis of classification will become pre-eminent.
The current classification largely relies on recognition of unique clinical phenotypes for differentiation of the various headache syndromes. However, some caution has to be exercised in classifying headache syndromes using this approach. Different headache syndromes can present with the same clinical phenotype; for example, there is a considerable overlap in the clinical presentation of SUNCT syndrome and ophthalmic trigeminal neuralgia, thereby often making it difficult to differentiate them in clinical practice (12). On the other hand, a headache syndrome may express different clinical phenotypes, e.g. migraine patients often also have phenotypic tension-type headaches, the biological basis of which is probably different from pure tension-type headaches; certainly treatment response is different (13).
The reliance on a specific treatment response to classify medical disorders is more controversial. The current classification criteria of the International Headache Society (IHS) require an absolute response to indomethacin for the diagnosis of PH and hemicrania continua (HC) (14). However, this is problematic for several reasons (15). First, the therapeutic response in patients with PH or HC is not exclusive to indomethacin and a variety of other medications have been reported to be effective. Patients with PH or HC would not receive the correct diagnosis in the event that a response to another medication occurred prior to a trial of indomethacin. Second, other primary headache syndromes, such as cough headache, primary stabbing headache, hypnic headache and sexual headache, also often respond in a highly consistent manner to indomethacin. A response to indomethacin is therefore not specific to PH or HC. Finally, patients have been described who had the clinical phenotype of PH or HC but did not respond to indomethacin. This raises the question whether there is a subset of patients with the underlying biology and clinical phenotype of these syndromes who do not respond to indomethacin. If the mode of action of indomethacin involves interrupting the central pathogenic mechanism of these syndromes, then it is likely that all patients will respond to indomethacin and the indomethacin-resistant cases do not represent true PH or HC. These considerations highlight the tenuousness of including specific treatment response as a diagnostic criterion. Indeed, some authors have suggested that the classification of primary headache disorders, in the absence of a genetic or biochemical marker, should be based on clinical features alone and should remain independent of treatment response (16). Nonetheless, the IHS classification committee have taken the uncomfortable position of requiring an absolute response to indomethacin in the definition of PH and HC to homogenize clinical studies until the response is understood (14). It is hoped that this compromise position will be useful both in clinical practice and for research purposes. For clinical practice, it is prudent to diagnose PH or HC only in patients who are responsive to indomethacin. For research purposes, these syndromes need to be studied in their purest form, which would not be possible if the requirement for indomethacin responsiveness was discarded.
So what are the implications of these considerations for the classification of the cases purported to be bilateral PH? Trigeminal autonomic cephalgias, which include cluster headache, PH and SUNCT syndrome, are characterized by unilateral headaches occurring in association with cranial autonomic features (14). The clinical phenotype of the three case reports of bilateral headaches differs fundamentally from PH in lacking unilaterality and associated cranial autonomic features. We consider these differences to be cardinal and hypothesize that the bilateral headaches may represent a novel headache syndrome. We propose that bilateral, short-lasting, frequent headaches with the temporal profile of PH but lacking cranial autonomic features should be classified separately from PH and submit the term ‘bilateral paroxysmal cephalalgia’ to describe this novel phenotype. We encourage readers to report further cases of ‘bilateral paroxysmal cephalalgia’ and thereby broadened its clinical phenotype. We suggest that the term ‘bilateral paroxysmal cephalalgia’ should be limited to indomethacin-responsive cases until it is clear what the indomethacin response means. It is only with further studies of PH and ‘bilateral paroxysmal cephalalgia’ that it will be possible to elucidate the biological basis of these disorders and thereby determine whether they are variants of the same disorder or entirely different disorders.
Finally, the three case reports of ‘bilateral paroxysmal cephalalgia’ emphasized the importance of a trial of indomethacin in all patients who present with unilateral or bilateral, short-lasting, frequent headaches, with or without cranial autonomic features.