Abstract
We report a patient with a secondary form of chronic cluster headache, caused by an intracranial presentation of inflammatory myofibroblastic pseudotumour located in the posterior fossa, with total remission of the pain after resection. The headaches were resistant to many of the usual treatments for cluster headache. The patient had two normal computed tomography scans and one normal magnetic resonance imaging of the head before the additional diagnosis of brain tumour was made. This is an unusual cause of cluster headache with intracranial mass, with an unexpected clinical presentation, a rare triggering manoeuvre, unusual pathology and successful treatment. This patient probably had the hypothalamic biological predisposition to cluster headache and, when a small mass disturbed pain-sensitive structures in the posterior fossa, it excited the trigeminovascular system via posterior fossa trigeminal and upper cervical afferents, and triggered the pathophysiological processes that resulted in a secondary form of chronic cluster headache.
Keywords
Introduction
Cluster headache (CH) is a rare form of intermittent, short-lived excruciating unilateral head pain associated with autonomic dysfunction, with a prevalence rate ranging from 0.05% to 2.4%(1–3). The Ad Hoc Committee on the Classification of Headache, in 1963, classified it as a primary headache disorder (4). In 1988, the Headache Classification Committee of the International Headache Society (5) defined the diagnostic criteria of CH, and established two subdivisions, episodic and chronic.
In spite of the well-defined diagnostic criteria and the characterization of CH as a primary headache with no known cause, in 3–5% of patients this syndrome is secondary to a diverse variety of intracranial structural abnormalities (6–10).
We present a patient with features of chronic CH probably secondary to a rare form of structural lesion located in an intracranial region not usually associated with these symptoms. The headaches were resistant to many of the usual treatments of CH, and the patient had complete resolution of his headaches after surgery.
Case report
A 22-year-old man came to our attention 6 years ago, presenting with a 3.5-year history of headache. He was a violin student whose pain prevented him from keeping up with his practising and study. He described the pain as severe, sharp, steady but sometimes pulsating and located around the left eye and temple. Attacks began with a frequency of twice a day, but by the time of his first visit the frequency was four times a day, with an average duration of 2.5 h. During the attacks there was ipsilateral lacrimation and conjunctival injection and the patient remained agitated. Sometimes there was ptosis. The attacks occurred both during sleep and throughout the day, usually in the morning and afternoon. When he played his violin sometimes the requisite neck position triggered the pain. The patient does not usually drink alcoholic beverages and alcohol could not be defined as a potential trigger. Past medical and family histories were unremarkable. Patient did not have a previous history of head trauma. None of his relatives had CH or migraine. The physical and neurological examinations in the attack-free periods were normal. Routine haematology and blood chemistry studies were also normal, as were two contrasted computed tomography (CT) scans of the brain, performed 1 and 2 years before coming to us. Our diagnosis was primary chronic CH and we ordered a magnetic resonance imaging (MRI) with gadolinium which was unremarkable. Different therapeutic approaches were tried throughout the following 6 years. The usual pharmacological preventive treatments for CH, such as high-dose verapamil, daily ergotamine tartrate, lithium, divalproex sodium and other anti-convulsants provided no relief. A 1-week hospitalization on an interdisciplinary in-patient headache unit with various treatments including intravenous repetitive dihydroergotamine was not helpful. Oxygen provided acute moderate although transient relief. Indomethacin in a sustained relief preparation, at a dose of 75 mg bid (to a maximum dose of 75 mg tid), provided reasonable but only partial relief. Blockade of the occipital nerves and cervical trigger points was followed by significant although transient relief.
Two years later he noted an exacerbation of his clinical picture, and he described an aching at the base of the skull on the left side with radiation of discomfort into the adjacent temple. Holding the violin on his left shoulder and tilting his head typically exacerbated the headaches. At the patient's suggestion a repeat MRI was performed and showed a 1.8-cm contrast enhancing mass in the left posterior fossa, which appeared to be attached to the undersurface of the tentorium and was felt to have all the characteristics of a meningioma (Fig. 1). There was no mass effect. The situation was discussed with the patient, and he was told that the tumour could enlarge with time, and because of his young age surgery was the suggested option for treatment. He was told the surgery was not likely to affect his headache. Surgery was performed with complete tumour removal as shown on the postoperative MRI scan (Fig. 2). The preliminary pathology report was not conclusive, despite numerous immunohistochemical stains, and the initial diagnosis was pseudotumour, not meningioma. Two neuropathologists concluded, after immunohistochemistry slide reviews, that the mass was benign and consisted of an inflamed mesenchymal nodule, with the final diagnosis ‘inflammatory myofibroblastic pseudotumour’ (Fig. 3).
Pre-operative magnetic resonance imaging showing dense mass adhered to the undersurface of the left tentorium.
Post-operative magnetic resonance imaging demonstrating complete tumour removal.
Histological microphotography of the inflammatory myofibroblastic pseudotumour.
Following the immediate postoperative period, the patient reported complete relief of his headaches. He had been taking indomethacin, mirtazapine and nefazodone (at that time used for the psychological symptomatology of the patient) for an extended period of time. Two months later the mirtazepine was tapered and stopped, followed by the indomethacin and then nefazodone. Since the surgery he has not had a single cluster headache (19 months of follow-up). He is on no medication and works regularly, playing the violin without consequence.
Discussion
Cluster headache is a clinically well-defined disorder in which patients suffer extremely painful headaches with clock-like regularity, with or without a period of remission, associated with autonomic dysfunction (11). The three major pathophysiological aspects of CH are as follows: (i) the trigeminal distribution of the pain, more often in the first two divisions than in the third; (ii) the associated autonomic features; (iii) the clock-like episodic pattern of the attacks with circadian and, at times, circannual rhythmicity. There is evidence locating the site of pathology of this disease in the periphery of the nervous system (12, 13), but recent research suggests that CH is a central nervous system disorder involving the posterior hypothalamic circadian cycling mechanisms (14). May et al. (15) showed activation of the hypothalamic grey on PET scanning in CH, this area being of obvious interest because of its role in the control of circadian rhythms (16).
Sometimes a secondary form of CH follows cranial or vascular abnormalities, and is clinically indistinguishable from the primary form and may even respond to the typical CH medications (4). Temporal arteritis, sinusitis, glaucoma, intracranial aneurysms, tumours, arteriovenous malformations, dissection of the cervicocephalic cerebral blood vessels, and even cervical cord lesions or infarctions have been known to mimic CH (17, 18). CH is sometimes triggered by manipulation of the trigeminal nerve, such as after dental procedures or sinus surgery.
We have presented a case of secondary chronic CH with some interesting aspects: (i) the patient had CH at times triggered by certain neck positions; (ii) two CT scans and one MRI of the brain, done over several years, were normal prior to the diagnosis of posterior fossa mass; (iii) the last MRI demonstrated a lesion not expected to be causing the pain; (iv) blockade of the greater occipital nerves and injections into upper cervical trigger points were followed by significant relief; (v) the surgical extirpation of the lesion was followed by immediate and total relief from CH; (vi) the lesion was a rare type of focal, nodular, inflammatory myofibroblastic pseudotumour. It appears certain that this mass in the posterior fossa was at least a major trigger factor for, and maybe even causative of, the patient's pain (pain elicited in certain neck positions, relief with blockade of occipital nerves and injections into cervical trigger points and total relief after the surgery).
Inflammatory pseudotumour is an uncommon and enigmatic lesion. The spindle cells found in this tumour have features of myofibroblasts. Because of the unclear relationship of these lesions with inflammatory fibrosarcoma and their indefinite biological behaviour, inflammatory pseudotumour is currently classified as inflammatory myofibroblastic tumour (IMT). To date, only case reports or small series have been published about these tumours (19). IMT is the most frequent primary tumour of the lung in childhood, sometimes having an extrapulmonary location, but it is infrequent in the brains of children or adults (20). As far as we know, this is the first case of headache as the primary manifestation of IMT reported to the literature. A review of 84 cases of IMT occurring in extrapulmonary sites of involvement showed the following distribution: abdomen, retroperitoneal space, or pelvis (n = 61); head and neck, including upper respiratory tract (n = 12); trunk (n = 8); and extremities (n = 3) (21). Its typical clinical presentation includes mass, fever, weight loss, pain, and site-specific symptoms. Our patient had only site-specific pain.
Occasionally posterior fossa lesions, such as meningiomas (10), epidermoid tumours (22), vertebral artery aneurysms (23), etc., have been reported to cause secondary CH, eliciting anatomical and pathophysiological discussion.
As reviewed by Hardebo (24), the majority of pain fibres of the head, including those in cranial vessels, travel within the three divisions of the trigeminal nerve on their way to synapse in the spinal trigeminal nucleus; the cell bodies of these primary efferent neurones are located in the Gasserian ganglion. Pain fibres from the occipital region and ear, the dura mater of the posterior cranial fossa, and probably the vertebro-basilar arterial tree, travel via 2nd and 3rd cervical spinal ganglia to synapse in the spinal trigeminal nucleus (25–27). The surface of the tentorium, on the other hand, is innervated by V1 above the tent and C1 and C2 below, from a recurrent tentorial branch described by F. McNaughton from the Montreal Neurological Institute (A. Purdy, personal communication, from (28)). There is no surface or skin representation of C1, so the sensitive input is highly referred to V1 and vice versa. Also in the trigeminal nucleus, V1 is juxtaposed to C1 in the dorsal root entry zone (A. Purdy, personal communication, from (29)). This could open possibilities for referred pain – with pain felt in the neck and ear when actually being generated from the trigeminal distribution, and vice versa, due to convergence upon the same pool, or spill over to the neighbouring pools, of secondary neurones at the location of the spinal trigeminal nucleus (26). Involvement of upper cervical roots may thus give rise to symptoms of CH, that are relieved upon cervical nerve blockade or decompression (30).
It is well known that extreme ophthalmic division pain of many types can trigger flow changes in the cavernous sinus as part of the trigeminal–parasympathetic reflex (29), elicited by connections between the trigeminal nucleus and the sphenopalatine ganglion, via the superior salivatory nucleus (31). The sphenopalatine ganglion is located deep in the pterygopalatine fossa. Its parasympathetic roots are implicated in the autonomic dysfunction seen in CH; one of its branches is sometimes sectioned (the greater superficial petrosal nerve) as a surgical option for the treatment of cluster (31).
The key pathologicalprocess that takes place in CH appears to be central (32). The chronobiological disorder observed in this syndrome could be explained by the involvement of the suprachiasmatic nucleus of the hypothalamic region. It is entirely plausible, however, that some central permissive process entrains or releases the trigeminovascular pain system. We present a case that could illustrate this situation. Perhaps our patient has the biological predisposition to CH and, when a small and very unusual mass disturbed pain-sensitive structures in the posterior fossa in the distribution of the upper cervical nerves, worsened by the patient's requisite neck position while playing the violin, it triggered the pathophysiological process discussed above.
Two cautionary notes must be proffered: first, it is possible, though most unlikely, that the patient went into spontaneous remission of his chronic CH after so many years. Against this possibility, we can argue: (i) the patient never had one remission in the 9.5 years since his pain began; (ii) the patient is pain free without any medication in the last 19 months of follow-up; (iii) absolute remission of cluster-type pain was described as early as the first postoperative day. In addition, since the patient had a mild response to indomethacin, the diagnosis of chronic paroxysmal hemicrania (CPH) could be considered. Again, three arguments against this possibility can be made: (i) the duration of pain per attack was clearly too long; (ii) the number of attacks per day were much too few; (iii) the response to indomethacin was not complete. According to the International Headache Society Classification (5), the diagnosis of CPH cannot be established.