Abstract
The demonstration by magnetic resonance imaging (MRI) scanning of thickening and enhancement of the cisternal part of the oculomotor nerve in patients diagnosed as ‘ophthalmoplegic migraine’ prompts reconsideration of this uncommon disorder. The case histories of five patients, three male and two female, varying in age from 6 to 30 years, are presented here. Recurrent painful ophthalmoplegia started in infancy in two cases, childhood in two instances and adult life in one. One child had his first attacks at 3, 5 and 12 months of age, on each occasion 10 days after an injection of triple vaccine. The possibility of this condition being a recurrent demyelinating neuropathy is considered and its possible relationship to migraine explored.
Introduction
Pearce (1) attributes the first account of the association of recurrent headaches with ophthalmoplegia to Gubler in 1860 and cites Charcot's 1890 paper ‘sur un cas de migraine ophtalmoplegique’, which gave a name to the syndrome. Before angiography and computed tomography (CT) scanning, compression of the oculomotor nerve by aneurysm or other causes could not be excluded as simulating this uncommon disorder. In 1960, Walsh and O'Doherty (2) postulated that oedema of the wall of the internal carotid artery generated during migraine headache could compress the third, fourth or sixth cranial nerves in the carotid sinus. Vijayan (3) pointed out that pupillomotor fibres were usually affected early by compression of the third cranial nerve, which was not always the case in ophthalmoplegic migraine. He considered that ischaemia of the nerve, as in diabetic neuropathy, was a more likely explanation.
The similarity of the symptoms to those of Tolosa–Hunt syndrome has encouraged speculation that ophthalmoplegic migraine may be a variant of this disorder (4), but the lack of involvement of the ophthalmic division of the trigeminal nerve and the recent demonstration of gadolinium uptake within the paretic oculomotor nerve (5–11) clearly differentiate it from a granuloma in the cavernous sinus compressing the nerves within it (12, 13). A definitive study by Mark et al. (10) comprises six patients studied by magnetic resonance imaging (MRI) in the acute phase who all showed intraneural enhancement with enlargement of the oculomotor nerve at the point where it emerges from the midbrain and extending along the proximal course of the nerve. The MRI findings in reported cases and one of our own prompted us to report our limited experience of ophthalmoplegic migraine and to question whether the condition should be reclassified as a recurrent neuralgia, possibly a demyelinating neuropathy.
Case resorts
Case 1
ML, a girl aged 16 years, had been subject to attacks of screaming, associated with drooping of her left eyelid, since the age of 9 months. The episodes recurred about twice a year, lasting from 2 to 10 days with an association between headache and left ptosis becoming clear from the age of 3 years. In the 12 months before consultation she had experienced eight attacks, commonly lasting 2–4 days but occasionally extending to 10 or 11 days. Each episode started with a sharp pain behind the left eye, fluctuating in intensity, followed after 3 or 4 days by drooping of her left eyelid. She became nauseated, sometimes vomited and was sensitive to light, sound and smells. Her parents considered that these attacks were brought on by stress, fatigue, heat or glare. Cafergot tablets and aspirin taken at the onset of symptoms had sometimes eased the headache. She had not been helped by amitriptyline or pizotifen as prophylactic therapy. Her mother and maternal grandmother had suffered from migraine. Unenhanced MRIs of the brain at the ages of 12 and 14 years were normal.
She was first examined when an attack was in progress. She had a left ptosis with paresis of upward deviation in the left eye which turned outwards as she attempted to look up. Her left pupil was slightly larger than the right but reacted to light. Because of the severity of her headache she was admitted to hospital. MR angiography was normal. No enhancement was seen in the oculomotor nerve at the time of vascular imaging but delayed films showed marked uptake of gadolinium at the point of exit of the nerve from the midbrain, continuing along the line of the nerve (Fig. 1).
Magnetic resonance imaging showing enhancement of the intracisternal portion of the oculomotor nerve with gadolinium in a case of ophthalmoplegic migraine. (a) Coronal image. (b) Axial image.
Her headache persisted while her oculomotor palsy progressed until it became complete after 3 days with a dilated left pupil unresponsive to light. It started to resolve 2 days later as her headache subsided. When she was reviewed 6 weeks after the onset, her ptosis had cleared up and adduction of her left eye was full. Elevation of the eye was still impaired and the left pupil was slightly larger than the right although responsive to light. A repeat MRI scan then showed that enhancement of the oculomotor nerve was still present but less intense. Her ocular symptoms and signs had disappeared completely 2 months after the onset.
She remained free of headaches for 7 months while taking flunarizine 10 mg each night, then had two episodes of headaches lasting 36 h without ocular signs when this medication was stopped. She has not had any other episode in the 4 months since dosage was resumed as 5 mg each night.
Case 2
The story of JP, a 9-year-old boy, started with three episodes in which his right eyelid drooped and his right eye turned out at the ages of 3, 5 and 12 months, respectively, each following 10 days after an injection of triple vaccine. These attacks were associated with vomiting and apparent headache. From the age of 2 years he started to have attacks every 6 months but the frequency increased to once every 2 months at the age of 7 years.
He usually awoke with pain above and behind his right eye associated with photophobia, nausea and occasional vomiting. If he was given an analgesic at the onset, he often went to sleep and awakened without headache. If the headache persisted, his right eye turned outwards and his eyelid drooped on the second day (Fig. 2). His mother had experienced two attacks of severe migraine headache in her life. A follow-up telephone call when he was aged 19 ascertained that most ocular symptoms and signs subsided at the age of 10 years, but that he continued to have two or three headaches each month, often accompanied by dilatation of his right pupil. No investigations had been undertaken.
Case 2 at 2 years of age showing a right ptosis persisting after headache ceased. Photograph kindly provided by his parents.
Case 3
SS was seen when he was aged 6 years. Just after he turned 5 years, he developed a right-sided headache that continued for 4 days, accompanied by vomiting, before his right eyelid drooped and he had difficulty in looking upwards and inwards with his right eye. His mother said that his right pupil was larger than the left. A CT angiogram, unenhanced MRI of the brain and CSF study were all normal. His headache resolved after a week and his eye movements gradually recovered over another 10 days.
Since then he has had headaches every 4–6 weeks which were mid-frontal, associated with vomiting and sensitivity to sound, persisting for 6–48 h. During one episode he complained of double vision and his right eyelid drooped. Some headaches were preceded by his seeing dots in front of his eyes. He did not respond to pizotifen, imipramine or sodium valproate but headaches were shortened by ibuprofen and a medihaler of ergotamine if used at the onset. The frequency of attacks reduced on cyproheptadine to once every 8 weeks.
Two and a half years after his first attack he complained of intermittent loss of vision and an ophthalmologist found bilateral papilloedema. A CT scan of the brain was normal and lumbar puncture disclosed an opening pressure of 28 cm. He was treated for benign (idiopathic) intracranial hypertension with acetazolamide, followed by a low daily dose of prednisone. His papilloedema resolved over the next 3 months but he continues to have migrainous headaches without ocular symptoms at a lower frequency.
His mother had had two attacks of migraine in her life.
Case 4
CS, a 13-year-old boy, had suffered from left frontal headaches since the age of 3 years, recurring every 6 weeks or so, lasting for 1–3 days, accompanied by copious vomiting and photophobia. Two months before consultation he developed double vision on the second day of a headache when he was noted to have impairment of the third and possibly sixth cranial nerves on the left with sparing of the pupil. This took 3 weeks to resolve. Sixteen months later one of his headaches was associated with ptosis. CT, unenhanced MRI and MRA scans were normal. His maternal grandmother had been subject to migraine.
When last contacted at the age of 19 years, he continued to have headaches every 2 months but without any ocular symptoms.
Case 5
LN, a woman aged 30 years, had experienced right frontal headaches since childhood, recurring every 4–6 weeks, lasting 2–3 days on each occasion. They were preceded by yawning, were pulsating in quality and were accompanied by photophobia and slight drooping of her right eyelid. The day before consultation she had woken up with a right retroorbital and frontal headache. Later that day, her right eyelid drooped and she had double vision. When examined the following day, she had a moderate ptosis and inability to elevate and adduct the right eye. Her right pupil was larger than the left and reacted poorly to light and accommodation. There was no impairment of facial sensation and the corneal reflexes were brisk. CT and unenhanced MRI scans of the brain and MR angiography were normal. One of her sisters, but not her parents, had suffered from migraine. She moved to another country shortly after the episode described above and contact with her has been lost.
Discussion
Headache and recurrent oculomotor palsy started in childhood in four of our five patients. One boy had only two episodes of ophthalmoplegia in his early teens but continued to have migraine headaches thereafter. Eye signs followed the onset of headache by 12 hours to 4 days and persisted for up to 2 months. Follow-up MRI in the cases reported by Mark et al. (10) demonstrated resolution of oculomotor nerve enhancement in 7–9 weeks. Persistence of the MRI appearance for 40 days was reported by Prats et al. (11) in one boy with repeated episodes. In their four patients the MRI was abnormal in two out of the three in whom it was examined and was found to be normal 2 and 4 years later, respectively, when the patients were suffering episodic migraine attacks without ophthalmoplegia. There is no reference to gadolinium being injected in the one patient in whom the MRI was negative.
There were unusual features in two of the patients reported here. The linkage of three episodes in Case 2 with injections of triple vaccine has not been recorded before and suggests the possibility of an immune reaction affecting the oculomotor nerve. The development of benign intracranial hypertension in Case 3 brings up the possibility that the ophthalmoplegic episode more than 2 years previously might have been caused by an unrecognized increase in intracranial pressure. This is considered unlikely because he did not have papilloedema at that time and lumbar puncture findings were recorded as normal. Whether benign intracranial hypertension could be related aetiologically to the earlier ophthalmoplegic episodes remains uncertain.
What is the mechanism of ‘ophthalmoplegic migraine’ and how should the syndrome be classified? Although the recurrence of ocular palsies and unilateral headache bears comparison with Tolosa–Hunt syndrome, the absence of a demonstrable lesion in the cavernous sinus by CT and MRI (12, 13) and the presence of marked uptake limited to the intracisternal portion of the oculomotor nerve make the distinction between the two syndromes quite clear. Mark et al. (10) provide a comprehensive differential diagnosis of the MRI finding of enhancement of the oculomotor nerve which may appear in HIV and other infections as well as in demyelinating conditions such as Miller Fisher syndrome. Such uptake is seen in the facial nerve in Bell's palsy (14) and in idiopathic trigeminal sensory neuropathy (15), but not in ischaemia of the oculomotor nerve in diabetic neuropathy (16), even though the latter is often associated with severe unilateral headache (17). Lee et al. (18) reported a gadolinium-enhanced lesion in the line of the sixth cranial nerve course within the brainstem in a patient with ophthalmoplegic migraine affecting this nerve.
The MRI appearances therefore favour demyelination or inflammation of the oculomotor (or occasionally the trochlear or abducent) nerve in cases of ‘ophthalmoplegic migraine’, rather than an ischaemic lesion. The most striking feature of the usual MRI changes is the globoid appearance of the first part of the oculomotor nerve on emerging from the midbrain, indicating a gross localized enlargement of the nerve, consistent with intraneural oedema as reported in some experimental demyelinating neuropathies (19) and in chronic inflammatory demyelinating polyneuropathy (CIDP) (20). Arroya and Horton (21) reported that a patient with recurrent CIDP developed a right periorbital pain associated with an oculomotor palsy which resolved after 1 month. MR scan was normal but there is no record of the scan being enhanced by gadolinium. These authors cite the frequency of ocular motor involvement as being 3–4% of patients with CIDP. Waddy et al. (22) found a lateral rectus palsy in three of their five cases of CIDP and a fourth had associated ptosis and diplopia. Such ocular palsies may precede chronic relapsing polyneuropathy by periods of 19 days to 3.5 months (23).
Mark et al. (10) considered that a benign viral or idiopathic inflammatory neuropathy was the most likely underlying cause in most patients with ‘ophthalmoplegic migraine’. We would favour the latter because recurrent CIDP is well recognized and ‘in rare patients there is a history of preceding infection, immunization or receipt of biologic materials (sting or injection) within a few weeks of the onset of symptoms or associated with relapses’ (24). The second patient in our series with relapses on three occasions 10 days after immunization with triple vaccine supports this concept. Moreover, pain is a common feature in CIDP (25) and also in the Guillain–Barré syndrome (GBS) (26), particularly in children (27). Enhancement of the nerve roots with gadolinium on MRI is also common in CIDP (20) and GBS (28), and correlates with both the presence and degree of pain (28).
If ‘ophthalmoplegic migraine’ is a recurrent neuritis or demyelinating neuropathy, what is its relationship to migraine headache? In some instances ptosis and diplopia, or both, become superimposed on a fairly typical long-standing history of migraine, affecting the oculomotor nerve on the habitual side of headache (e.g. Cases 4 and 5). On the other hand, some patients start with unilateral headache followed by oculomotor palsy, suggesting that the primary event may have been an immune response in the nerve, possibly vasculitis, causing neuralgic pain on which some migrainous features such as nausea and photophobia develop (e.g. Cases 1–3). It is known that the human oculomotor nerve contains sensory ganglion cells and it has also been demonstrated in primates and other species that sensory fibres from the ophthalmic division of the trigeminal nerve enter the oculomotor nerve, pass through it into the brainstem and terminate in the spinal trigeminal nucleus (29). It is possible that an inflammatory process affecting the oculomotor nerve could irritate these trigeminal sensory fibres and subsequently activate the trigemino-vascular system in people who are already migraineurs, thereby triggering the associated migrainous headaches. Repeated MRI scans with gadolinium in periods of freedom, and periods of headache before and after ophthalmoplegia appears should serve to clarify the sequence of events if not the mechanism.