Abstract
The present study summarizes the authors' experience of the clinical profile of short-lasting trigeminal autonomic cephalgias (TAC) in Indian patients. Over a period of 17 years a total of 41 cases of episodic cluster headache, seven cases of chronic cluster headache, six cases of variant cluster headache, three cases of paroxysomal hemicrania, and a single case of SUNCT syndrome were encountered. TACs appear to be rare in Indian patients and cluster headache seems to be exclusively a disease of men. The present report is to the best of our knowledge the first of its kind to be reported from India.
Introduction
The trigeminal autonomic cephalgias (TACs) are a group of primary headache disorders characterized by unilateral trigeminal distribution pain that occurs in association with prominent ipsilateral cranial autonomic features (1). The group comprises cluster headache, paroxysmal hemicrania and short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT syndrome). They must be differentiated from secondary TACs, and other short-lasting primary headaches, and then from each other. The concept of a short-lasting headache is naturally somewhat artificial in terms of defining ‘short’. A typical attack time is less than 4 h, in contrast with the major differential diagnosis of migraine in which attacks are usually longer. The differentiation between TACs is essential since the treatments are very different.
Cluster headache (CH) is an uncommon form of short-lasting primary headache syndrome with autonomic features. This is a remarkable condition in many ways, not least for the galaxy of appellations, eponyms and synonyms which it has attracted. First described in Gerhard Van Swieten's medical text book (2), earlier descriptions of the condition include those of Harris (3), Horton et al. (4), Kunkle et al. (5) and Symonds (6). More recent contributions include those of Ekbom (7), Pearce (8), Kudrow (9) and Manzoni (10). The clinical profile has been recently reviewed by Nappi and Russel (11), Newmen et al. (12) and Matharu and Goadsby (13). Cluster headache occurs in two forms, episodic and chronic. The diagnostic criteria for these have been laid down in the International Headache Society Classification 1988 (14), and in the proposed draft revision (15).
Paroxysmal hemicrania (PH), like CH, is characterized by strictly unilateral brief, excruciating headaches that occur in association with cranial autonomic features. PH differs from CH mainly in the higher frequency and shorter duration of indivudual attacks, though there is a considerable overlap in these characteristics. However, unlike CH, PH responds in a dramatic and absolute fashion to indomethacin, thereby underlining the importance of distinguishing it from CH. PH has both episodic and chronic forms that will be defined in exactly the same way as is done for CH in the revision of the IHS classification (15).
SUNCT, like other trigeminal autonomic cephalgias, manifests as a unilateral headache that occurs in association with cranial autonomic features. The features that distinguish it from other trigeminal autonomic cephalgias are: very brief duration of attacks that occur very frequently, and the presence of prominent conjunctival injection and lacrimation, both of which are present in patients.
The present report appears to be the first of its kind considering the clinical profile of short-lasting TACs (including variant forms) in Indian patients in a hospital-based prospective study. Certain parts of the study had been published earlier (16, 17).
Materials and methods
Between the years 1985 and 2002, in a Neurology out-patient clinic (tertiary care) of a busy general hospital, a total of 41 cases of episodic CH, seven cases of chronic CH, seven cases of variant CH, three cases of PH and a single case of SUNCT syndrome were seen. This prospective study was initiated in 1985 with the object of studying the clinical pattern of CH and its variants in Indian patients. The study title was recently changed to Trigeminal Autonomic Cephalgias, a collective terminology and one which is being increasingly used in the current medical literature.
The diagnosis of CH was established as per criteria of the International Headache Society, 1988 (14). Diagnosis of CH in cases seen between 1985 and 1988 (n = 3) was initially made based on standard textbook descriptions of attacks. However, subsequently on review of the case notes, all conformed to the diagnostic criteria laid down by the IHS. Diagnosis of PH was made as per suggested criteria proposed by Goadsby and Lipton (1). Diagnosis of SUNCT syndrome was based on the criteria laid down in the available literature and mentioned in the draft revised version of the IHS classification (http://www.i-h-s.org;2002). All patients underwent detailed clinical evaluation including maintenance of a headache diary. Routine haematological tests, biochemical tests and chest radiography were done in all as well as brain imaging by computed tomography or magnetic resonance imaging (MRI). Therapy responses were recorded carefully but will not be reported here (for CH cases) and will form part of a separate communication. The clinical profile of the patients studied is presented.
Results
Cluster headache
Tables 1 and 2 summarize the clinical profile of the 41 episodic CH and seven chronic CH patients, respectively, encountered during the study period. In the present study, both forms of CH involved exclusively male sex with onset age in late 20s or early 30s in most. Missed diagnoses were common. The right side was affected commonly with pain of boring nature in the orbito-frontal region. Nocturnal attacks were more common. Classical associated symptoms were commonly seen.
Clinical profile of episodic cluster headache patients (1985–2002) (N = 41; all men)
Clinical profile of episodic cluster headache patients (1985–2002) (N = 7; all men)
All patients had unremitting headache from onset and right-sided boring orbito-frontal pain.
All seven cases of chronic CH had been of the primary type, having unremitting headache spells from the onset for more than 2 years at presentation. All patients had been experiencing several attacks per week prior to therapy and noted similar associated symptoms as patients with the episodic form of the disease. However, none developed Horner's syndrome during an attack.
Neuroimaging studies excluded secondary causes of headache in all patients.
Cluster headache variants
This group consisted of six cases which had been difficult to classify phenomenologically and temporally in relation to frequency pattern and duration of headache. Two groups of patients were recognized.
Group 1
This consisted of four patients, all male with age range 25–37 years. At presentation they had episodic headaches (for 3–5 years) occurring in a ‘clustered’ form with daily attacks (one to two spells/day) occurring for periods of 2–4 weeks at onset. The duration of the ‘clustered’ periods increased in all as they continued to suffer, exceeding 8 weeks in all patients at presentation. Such ‘clustered’ headache spells were followed by headache-free periods lasting several months. The headaches had been either unilateral or bilateral, mostly temporal, throbbing in nature (not boring) associated with photophobia, phonophobia, restlessness and nausea/vomiting in most instances but unaccompanied by any aura, and fulfilled the diagnostic criteria of migraine without aura of the IHS classification. On most occasions in all four patients the headache would start around a fixed time in the middle of the night, waking them up from sleep. However, the headaches had never been associated with any autonomic nervous system (ANS) features and duration of headache spells exceeded 4 h. Graham (18) discussed similar patients along with typical CH. In our patients, the headache ‘clusters’ could be aborted with a tapering course of prednisolone (40 mg daily for 5 days and then taper in 5–7 days). Regular prophylaxis was not considered as the headache clusters were relatively infrequent and recurrences were controlled with further short courses of steroids.
The headache spells of these patients in isolation could be considered as migraine. However atypical features pointing similarity to CH were: (i) the ‘clustered’ nature of the attacks; (ii) occurrence of headaches mostly at a fixed hour of the day; and (iii) restlessness during the headache spells.
Termination of headache spells with prednisolone was considered significant and not related to the natural history, as in all cases the duration of the spells was much shortened compared with that at presentation.
However, an interesting observation was made in a single patient who could be followed up for a very long period. This patient, followed up between 1992 and 1998, has been re-examined recently (September 2003). During the years when he did not attend the clinic, he continued self-medication with prednisolone during headache spells. However, increasing doses and duration of therapy were needed. In 2002 his single headache spell continued for over 4 months and steroid did not provide much relief. In 2003, at presentation he had already been having headaches for more than 2 months, with no benefit with prednisolone 80 mg daily, and developed ipsilateral redness of eye and lacrimation during the headache spells. Thus he developed ANS features about 10 years after his headache onset. Improvement in headache severity and duration occurred with verapamil 240 mg daily.
Group 2
Two patients were included in this group based on their headache phenomenology.
Case 1 A 35-year-old male had been having frequent (one to two per week) episodic, mostly right temporal throbbing headaches, usually lasting > 4 h, without any ‘clustered’ pattern. He had photophobia during headache but no autonomic features. These headaches were partially controlled with flunarazine and β-blockers. For 1 year before presentation the pattern changed. Such throbbing headaches, lasting several hours, became associated with nasal stuffiness, lacrimation and extreme restlessness, but no redness of eyes or drooping. The headache severity and frequency were significantly improved with treatment with 240 mg verapamil daily in divided doses.
Case 2 A 45-year-old male presented with a 15-year history of spells (2–3 weeks) of daily headache occurring five to seven times every year. The headaches had always been throbbing in nature, shifting sides – right or left temporal – during each spell, lasted for 2–8 h and always accompanied by nasal blockage and ipsilateral lacrimation but no redness of eye or drooping. Significant benefit was observed with verapamil 240 mg daily. The unusual feature had been the long duration of many of the headache spells compared with classic description of CH.
Neuroimaging studies in all patients had been normal.
Paroxysmal hemicrania
Only three patients were encountered during the period of study.
Case 1
A 60-year-old lady presented with a few weeks’ history of sharp electric shock-like pain on the right side of the head starting from the nape of the neck and passing over the right ear up to the right temple. The jab would last for only a few seconds and was accompanied by a drop of tear coming out of the right eye but no redness or nasal blockage. She would suffer 50–60 such attacks daily. Neuroimaging and haematological investigations were normal. The pain was totally relieved after a course of roficoxib 50 mg daily for 2 weeks and did not recurr for the next 3 months of follow-up.
A diagnosis of episodic PH was preferred to occipital/upper cervical neuralgia in view of the absense of any background cervico-occipital pain and the presence of a prominant autonomic feature (tearing).
Roficoxib was preferred to indomethacin as the patient complained of ‘gastric trouble’ at presentation. Indomethacin responsiveness had not been considered essential for diagnosis of PH by Goadsby and Lipton (1) but was considered so in the current draft revision of the IHS criteria. COX-2 inhibitor responsiveness has of course been established in PH (vide infra).
Case 2
A 56-year-old lady presented with brief attacks of right temporal pain, throbbing in nature and lasting for 10–12 s. These had been occurring for the past 3 weeks and occurred two to three times daily when she would cry out. The pain was accompanied by watering, and a little redness of the right eye, but no nasal blockage or drooping. Investigations were normal. Near complete relief was obtained with indomethacin 100 mg daily, but pain recurred when the drug was withdrawn. The pattern of pain relief with indomethacin and recurrence on drug withdrawal continued for several months of follow-up (about 14 months). This case was diagnosed as chronic paroxysmal hemicrania.
Case 3
A 50-year-old lady presented with only a few weeks’ history of sharp electric shock-like pain from the nape of the neck to over the right ear to the right temple, accompanied only by tearing from the right eye. She would suffer about a dozen such pains daily. She had rheumatoid disease for over 20 years with peripheral joint deformities and was on methotrexate and prednisolone. She had equivocal pyramidal signs in all four limbs. Imaging studies revealed rheumatoid atlanto-axial dislocation. The pain, however, was relieved with indomethacin 100 mg daily within 3 days. She was sent for surgical treatment for atlanto-axial disease but did not return for follow-up.
Pain relief in the present case seemed unlikely to be related to the influence of indomethacin on the rheumatoid process, as a response was seen in only 3 days. Indomethacin responsiveness for the secondary PH pain seemed more likely.
SUNCT syndrome
A single case of SUNCT syndrome was encountered during the period of study and has already been reported in detail (17).
A 60-year-old male patient presented with approximately 12 years’ history of unremitting electric shock-like pain over the left side of the head. The pain would occur in bursts lasting for 4–6 min and occurred several times at 10–20-min intervals (over 100 times) during a day. Each burst consisted of several sharp jabs each of few seconds’ duration occurring at a few seconds’ interval, and affected the left side of the forehead and left periorbital region. This was associated with marked injection of the left eye, lacrimation and rhinorrhoea, nasal blockage and sweating over the forehead. A number of such pain spells were observed in the clinic, during which he would stop talking, sit quietly with eyes closed and the jaw slightly clenched. The left eyelids were oedematous, the left palpebral fissure narrower and the left eye looked injected. Pupillary size during and in between attacks was normal and equal and fundi were normal. There was no other neurological deficit. His MRI of the brain was normal.
Cold air seemed to be a triggering factor here, as many attacks were induced when he was made to sit in front of a pedestal fan in the clinic. There was no refractory period.
He had significant relief of pain with lamotrigine 200 mg daily but not with carbamezapine, indomethacin or topiramate.
Discussion
Cluster headache
The world-wide incidence or prevalence of CH is not known. Only one study has documented the prevalence to be 56/100 000 [95% confidence interval (CI) 31.3, 92.4] and incidence rate at 2.5/100 000 per year (95% CI 1.14, 4.75) in the Republic of San Marino (19).
CH seems to be an uncommon condition in India. Over a period of 17 years, only 48 cases were encountered by the author in a neurology out-patient clinic of a busy general hospital in a metropolitan city. Over a 1-year period (1989) in the same clinic only two patients with episodic CH were seen out of a total of 868 ‘new’ cases of primary headache disorder. Perhaps dedicated headache clinics would encounter more such cases. Although universally uncommon, CH appears to be much less common in India than in the west. Pearce (8) in the UK over a 12-year period encountered 101 such cases in a general hospital in the north-east of England. In this connection, there is a need for an accurate diagnosis. Most cases in the present series had a wrong referral diagnosis.
In India, at least judging from the present observation, CH seems to be a disease exclusively affecting the male sex. Only 10% of affected patients in an Italian series were women (10). The sex bias noted in Indian patients is worth noting. During an observation period of more than 30 years, Ekbom et al. (20) noted a trend towards a decreasing male preponderance; the male : female ratio was significantly higher among patients with onset before rather than after 1970. The sex bias noted in the present study may be related to environmental factors related to lifestyle pattern differences between Indian women and their western counterparts. The importance of lifestyle pattern in the epidemiology of CH has been mentioned by Finkel (21). A decreasing male : female ratio has been noted, perhaps attributable to lifestyle trends adopted by more women that were previously associated with men, such as tobacco use, alcohol consumption and working outside the home.
The present study was conducted in an institution which caters mostly for the middle-class Bengalee population of Calcutta and its neighbourhood. Tobacco use and alcohol consumption are still very uncommon amongst women in the Indian middle-class population. This factor combined with the overall rarity of CH in the population itself probably accounts for the striking sex bias noted in the present study. Ethnic and racial differences may be involved but difficult to comment upon in a hospital-based study. Referral bias seems unlikely to explain the sex difference noted in the present study, as in our hospital migraine population the sex difference noted is very similar to western figures.
The mean age of onset of symptoms in the current series is slightly higher than those reported in the west. It was 28.0 years in the series of Friedman and Mikropoulos (22), 27.5 years in the series of Ekbom (7), 29.6 years in that of Kudrow (9) and 28.9 years in that of Manzoni et al. (10). A family history of CH is rare and was not encountered in the present series. A family history of migraine was noted in 31.6% of cases. This is higher than what is expected to occur in the general population but similar to the observation of Pearce (8). Prodromal and premonitory symptoms have been described in the literature but were not observed in our patients. A few patients in the episodic CH group (n = 4) noted precipitation of attack by alcohol and avoided it during the cluster phase subsequently. Only one patient in the chronic CH group continued to consume alcohol and had attacks precipitated by it.
No seasonal variation of frequency of headache had been noted by our subjects. Kudrow (23) reported a relationship of frequency of attacks with daylight hours during a particular time of the year. Absence of such a fluctuation in our country may be related to less variation in daylight hours in India and other tropical countries compared with the western hemisphere. Most patients in the present series suffered from headache spells at night often waking them up. None experienced only day-time attacks, though some experienced both day and night-time attacks. Manzoni et al. (10) noted sharp peaks between 01.00 h and 02.00 h and between 13.00 h and 15.00 h. These timings probably represent resting hours of subjects. Taking afternoon naps is common in Italy and not uncommon in India, either. According to the diagnostic criteria of the IHS classification (14), each headache attack in CH should last not more than 3 h if untreated. In practice, with most patients in the present series, headache spells were much shorter (30–90 min). In Russel's prospective study (24) of 77 attacks, total duration was < 30 min in 29%, < 45 min in 62%, and < 1 h in 78% of patients. Such detailed prospective observation could not be made in the present study as most patients were on some form of medication (analgesics, ergots, migraine prophylactics) by the time they presented.
Most head pains were localized to the orbito-frontal region and in none did the pain spread to the malar region or jaw. Ekbom (7) commented on the lower syndrome with pain mostly at the upper jaw and malar region. Such a distinction apparently seems arbitrary and unnecessary. However, Cademartiri et al. (25) recently noted some differences in clinical features between the upper and lower syndromes and postulated involvement of different brain regions in the genesis of the pain in the two syndromes. The pain had been of boring type in almost all subjects, with occasional patients complaining of a burning pain or throb. Similar observations were made by Pearce (8). Almost all of the associated symptoms mentioned in the IHS criteria, except eye lid oedema, were noted in the present series. The frequency of occurrence of individual features corroborated those published earlier (11). It is noteworthy that recent contributions debate their role in diagnosis and physiopathology. In a large series of consecutive patients, Ekbom (26) found that in 3% of cases the diagnostic criteria were not met because of the absence of local autonomic signs during cluster attack. A similar observation was made by Nappi et al. (27).
The authors stress the importance of accurate diagnosis of this most fascinating disease. This can only be achieved by careful history elicitation, as doctors are unlikely to witness an actual attack in out-patient neurology practice.
Van Vilet et al. (28) in a nationwide survey in the Netherlands tried to analyse the factors involved in the diagnostic delay of CH. These include presence of photophobia or phonophobia, nausea, or episodic attack pattern and a low age of onset. Such features, no doubt, would suggest to the uninitiated a diagnosis of migraine and must have been implicated in the present series as well, where the majority presented with the referral diagnosis of migraine headaches.
Cluster headache variants
The six cases included in this category had been difficult to classify on the basis of headache phenomenology. Variant forms of CH have been mentioned earlier.
The cluster–migraine syndrome is diagnosed when elements of migraine headache occur simultaneously in patients suffering with CH, or vice versa. Solomon and Kappa (29) instituted arbitrary diagnostic criteria to establish the diagnosis of this uncommon syndrome. According to their criteria, the diagnosis of cluster–migraine syndrome was given to patients who had symptoms of one headache predominantly (either migraine or cluster), but in whom four or five features of the other headache were also present. Patients who experienced CH with nausea, vomiting, photophobia, or phonophobia would receive this diagnosis, as would patients with migraine who experienced ipsilateral autonomic features. The criteria employed by Solomon and Kappa (29) were not precise; the required number of associated symptoms was chosen arbitrarily. Other clinicians have reported patients with two distinct headache disorders. Graham (18) described patients who suffered from recurrent bouts of migraine headaches that recurred daily for days or weeks at a time, then entered a period of pain-free remissions. It is important to recognize the cluster–migraine syndrome because of the unique treatment strategies that need to be implemented. Patients suffering from this syndrome have been reported to respond to inhalation of 100% oxygen as an abortive strategy for acute attacks and lithium carbonate as prevention. This combination would not be expected to be helpful in patients suffering from typical migraine. Alternatively, β-blockers may occasionally help in this syndrome, whereas these agents would be unlikely to be useful in the treatment of CH.
The four patients included in group 1 indeed had fairly typical migraine headaches (without aura), but these occurred in a very periodic or ‘clustered’ manner as described by Graham (18), and were often associated with restlessness. They had no associated autonomic features at presentation. The spells could be terminated easily with short courses of corticosteroids, but this is not specific. Both migraine headaches and CH may benefit from corticosteroids. The response of this group of subjects to more specific CH medication like lithium or verapamil, however, was not tested initially in view of their infrequent headache spells.
One patient who could be followed up for a long period revealed interesting features. His ‘cluster’ periods increased over the years, the headaches became less and less steroid responsive and he developed ANS features with headache spells. Latterly, he responded to verapamil. This case is somewhat different from cases included in group 2, in that the patient's headache spells always occurred in a ‘clustered’ manner and lacked ANS features for a long period.
The two patients included in group 2 showed overlapping features of both migraine and CH (including autonomic features). However, all lacked the periodic or ‘clustered’ nature of the headache spells and resembled more chronic CH than episodic CH. They responded to specific CH medication such as verapamil but not to the same degree to antimigraine drugs. Case 2 probably had migraine headaches to start with on which CH features were subsequently superimposed. Such transition from migraine to other forms of TAC have been reported by Pearce et al. (30).
Paroxysmal hemicrania
PH is rare. Only three cases were seen during the study period. All three cases satisfy the diagnostic criteria suggested by Goadsby and Lipton (1) and included in the draft revised version of the IHS classification, except that indomethacin was not tried in case 1 (vide supra). Two of the patients responded to indomethacin which is thought near essential for diagnosis, and case 1 responded to rofecoxib. Response of PH to COX2 inhibitors has been reported in uncontrolled studies (31, 32). Case 1 can be designated as one of episodic PH and case 2 as one of primary chronic PH (not evolving from episodic PH). Case 3 deserves special mention. This lady had fairly typical PH symptoms relieved with indomethacin. However, she had significant rheumatoid atlanto-axial disease. It is likely that her symptoms resulted from involvement of the C2 root which carries autonomic fibres. This is secondary or symptomatic PH. Rapid response to indomethacin in this case would suggest indomethacin responsiveness of PH rather than effect of the drug on the rheumatoid disease process.
Cases of symptomatic paroxysmal hemicrania have been reported. Disorders that have mimicked paroxysmal hemicranias include aneurysms within the circle of Willis (33); arteriovenous malformations and cerebrovascular accidents (34); collagen vascular disease (35); Pancoast tumour (36); tunours of the frontal lobe (35), sella turcica (37), and cavernous sinus (38); intracranial hypertension (39); and thrombocythemia (33). A single post-traumatic case of chronic paroxysmal hemicrania (CPH) with typical migrainous aura has been described (40).
It may also be of interest that the pain distribution in case 1 and case 3 had been partly extratrigeminal. This was commented upon earlier by Dodick (41).
SUNCT syndrome
The present series includes the only published report of SUNCT syndrome from India (17).
The diagnosis of SUNCT syndrome in the present case had been clear. Trigeminal neuralgia could be excluded by the presence of significant autonomic features and failure to respond to carbamazepine. Chronic paroxysmal hemicrania seemed unlikely in view of non-responsiveness to indomethacin. Similarly, idiopathic stabbing headache was excluded by the presence of autonomic features and non-responsiveness to indomethacin.
The profile of the pain in the present case is slightly atypical in the sense that instead of individual jabs for a few seconds continuing at intervals, the patient had been experiencing spells of pain where there were recurrent jabs occurring in volleys. These pain spells were followed by a pain-free spell of several minutes (10–20 min) before the pain spell would return. Matharu et al. (42) described a patient who had repetitive jabs (like the present case) but ultimately developed a constant pain for 20–30 min.
The response to lamotrigine in the present case was striking. Lamotrigine has been reported to be highly efficacious in seven patients (42–45). Given in an open manner at a dosage of 100–200 mg/day, it induced complete remission in five patients and produced about 80% improvement in two patients. The ultimate confirmation of the utility of lamotrigine in treatment of SUNCT syndrome will have to await a double-blind placebo-controlled trial.