Unilateral Headache Attacks and Ipsilateral Atrophy of the Tongue Due to Neurovascular Compression of the Hypoglossal Nerve

The combination of unilateral frontal headache and acute ipsilateral hypoglossal palsy is typical of an infraclinoidal dissection of the internal carotid artery and is well known (1). Recently, the combination of acute severe parietal headache and idiopathic hypoglossal nerve palsy has also been described (2, 3). In a small case study, seven out of nine patients complained of localized, continuous, unilateral occipital pain (3). We have seen a patient who presented with acute-onset, repeated, unilateral headache attacks with severe atrophy of the ipsilateral tongue, probably due to neurovascular contact between the 12th cranial nerve and the right vertebral artery (VA).

Case report

In 2006 a 60-year-old man presented with strictly unilateral pain attacks on the right side that had started about 7 weeks before. The attacks lasted from several minutes to 1 h and occurred several times (minimally twice, maximally up to 10 times) during the day, but seldom at night. The pain was maximal in the occipital area with a radiation towards the temporal bone and right eye. Autonomic signs were not present during or between attacks. The attacks could not be provoked by any head or neck movement. At this time, although the patient could not remember if these symptoms also began parallel to the headache attacks, he noticed problems with the tongue, especially with tongue movements to the right side. The patient had pre-existing arterial hypertension, which was not regularly treated. Clinically, he was healthy, the only neurological sign being a paresis and a beginning of atrophy of the right side of the tongue.

Since the clinical symptoms also suggested atypical symptomatic paroxysmal hemicrania, we performed high-resolution magnetic resonance imaging (MRI) of the head, including MRI angiography. Surprisingly, clear neurovascular contact was found between the right hypoglossal nerve and the right VA, resulting in compression of the nerve at its root entry zone (Fig. 1). In addition, we also saw 70% stenosis of the right internal carotid artery and approximately 50% stenosis of the left internal carotid artery. No further abnormalities were detected. Electromyography of the tongue did not show spontaneous activity, but slight chronic neurogenic reorganization of the action potentials indicating longer-lasting alteration of the right hypoglossal nerve. Before the MRI scan, under the hypothesis of an atypical paroxysmal hemicrania, we began to attempt treatment with indomethacin (2 × 50 mg daily), which directly suppressed the attacks by > 90%. After a week the patient attempted cessation of the indomethacin medication, but after 1 day without the treatment the attacks reappeared and he immediately resumed the medication, which again suppressed the attacks almost completely. We did not subsequently switch to an anticonvulsive treatment, which is usually standard in neurovascular compression syndromes, since the patient was completely satisfied with the indomethacin treatment. Furthermore, we started treatment of his vascular risk factors with an angiotensin inhibitor, a statin and acetyl salicylate. More than 1.5 years later the patient is well on this medication, no further trial without the medication has been made and he has not complained of any subjective weakness of the tongue.

Figure 1

Neurovascular contact between the right vertebral artery (VA) and the right hypoglossal nerve (CN XII). (A) Axial 3D-FIESTA (fast imaging employing steady-state acquisition) magnetic resonance image demonstrating neurovascular contact between the right VA (arrow) and one nerve root bundle of the right CN XII (dotted arrow) at its point of exit from the brainstem. (B) Reconstructed, oblique plane, perpendicular to the CN XII nerve root bundle (for reconstruction plane see dotted line marked in A). Arrowhead, basilar artery; arrow, right VA; dotted arrow, CN XII nerve root bundle.

Discussion

We think that the development of occipital headache attacks with radiating pain towards the ipsilateral eye together with hypoglossal palsy supports the notion that there is a possible aetiological link between the two symptoms. The possible connection of hypoglossal nerve palsy and ipsilateral frontal or occipital-parietal headache has recently been described in three cases with so-called idiopathic reversible hypoglossal palsy (2). Otherwise, most publications focus on headache in combination with hypoglossal palsy due to dissection of the internal carotid artery and injury to the hypoglossal nerve extracranially in the region of the mandibular angle. In this situation the pain is normally located in the orbital region and is comparable to the known headache in carotid dissection without hypoglossal palsy (1). In our case, we believe that the hypoglossal palsy was due to neurovascular compression near the entry of the nerve in the brainstem, similar to the situation in classical trigeminal neuralgia. There are a small number of further reports on hypoglossal nerve palsies due to abnormalities of cerebral vessels (4, 5), most of which describe elongation of the vertebral artery (6) or tortuous dilation of the internal carotid artery (7). We have found only one case report that describes headache, tinnitus and tongue atrophy in a female patient with a dural arteriovenous fistula (8). These entrapment syndromes of the hypoglossal nerve are also supported by the anatomical post mortem study of Bademci and co-workers, who describe three possible contacts of the nerve with cerebral vessels (9). Otherwise, Yousry and co-workers were able to demonstrate contact between an artery and the hypoglossal nerve in about 61% of scanned healthy subjects (10).

A further question is how can a lesion of the motoric hypoglossal nerve cause a sensory symptom such as headache? First, it must be admitted that there is the possibility of random co-occurrence of the headache and hypoglossal nerve compression, especially as we do not know if both symptoms occurred together. Otherwise, some anatomical studies in rats have demonstrated that there are sensory fibres in the hypoglossal nerve that make contact with the lamina I neurons in the upper spinal cord (11, 12). Another study has shown that axon collaterals of spinal trigeminal neurons located in the pars caudalis and pars interpolaris project to the hypoglossal nerve (13). These results are also supported by the expression of c-fos in the lamina I-II of the upper spinal cord after electrical and chemical stimulation of the hypoglossal nerve (14). Thus, activation of these sensory fibres within the hypoglossal nerve as a result of the neurovascular contact should be able to elicit activation of upper spinal cord neurons and trigeminal neurons similar to the pattern discussed in the case of cervicogenic headache (for overview, see (15)). Indeed, the description of the pain localization in our patient resembles the pain in cervicogenic headache (15) or the attacks in so-called neck-tongue syndrome (16), with the exception that the attacks were too short and there was no provocation by head movement. Another explanation could be that there is evidence of postganglionic parasympathetic neurons in the trunk of the hypoglossal nerve, which express vasoactive intestinal peptide, muscarinic receptors and nicotinic receptors (17). Thus, compression of the hypoglossal nerve may be able to activate the trigeminal-parasympathetic reflexes active also in migraine and trigemino-autonomic headache syndromes. The obvious benefit which our patient received from indomethacin therapy is not easily explained, since it is still not known why indomethacin is effective in hemicrania continua and chronic paroxysmal hemicrania, and spontaneous remission of the attacks is possible. In contrast, it should be remembered that the patient attempted to cease the medication, and the attacks re-occurred instantaneously.