Deep brain stimulation of the posterior hypothalamic area in intractable short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT)

Case report

This 73-year-old man presented with a 6-year history of short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT). In 2003, he experienced right sided acute attacks of sharp stabbing pain in the 2nd and 3rd trigeminal dermatome lasting up to 30 s for the first time. There was no previous trauma. Initially, in a primary care centre he was diagnosed as having atypical trigeminal neuralgia and was pain free for about 6 months under a therapy with carbamazepine. In 2004, the attacks started again on an episodic basis. Since 2005 the attacks became more frequent with no attack-free intervals lasting longer than 4 weeks. Since onset, he noted a seasonal increase in attack frequency clustering in spring and autumn.

The attacks start with a fronto-temporal feeling of pins-and-needles 5 s before the stabbing pain emerges in the right-sided retro-orbital region with an occasional spread to the other trigeminal dermatomes. Attacks last 1–60 s. Maximal attack frequency was up to 30/h. There is no association with the circadian rhythm. Attacks were evenly distributed throughout the 24 h cycle including nocturnal attacks, frequently waking him up from sleep. Simultaneous to the pain, the patient experiences a facial autonomic activation with conjunctival injection, nasal and ocular lacrimation, a ptosis and occasionally a miosis. Attacks are strictly right-sided. Pain intensities vary between 8–10 points on the analogue scale. The patient did not notice any trigger zones, such as during teeth brushing, food intake, swallowing, shaving, touching the oral mucosa or drinking cold fluids. The only occasional trigger noted was quickly getting up. As he stopped drinking alcohol after the emergence of the attacks, he could not make a statement about alcohol sensitivity.

The patient’s condition failed to improve under previous medications including indomethacin, pirinitramide, azathioprin, intravenous and oral corticosteroids, ipsilateral occipital nerve block, pregabalin, mirtazapine, oxygen, triptans, lorazepam, tramadol, gabapentin, baclofen, phenytoin and oxcarbazepine. In 2004, the patient exhibited signs of a carbamazepine intoxication and developed a sick sinus syndrome with a 2–3° sinoatrial block making a permanent application of a DDD-R cardiac pacer necessary. Frequent elevations of liver transaminases led to multiple changes in medication. Since 2005, the patient was in our continuous care with multiple admissions due to exacerbation of pain or side-effects of medication. Other medical disorders include Dupuytren's disease, diabetes mellitus type II and arterial hypertension.

Previous diagnostic work-up included an MRI and two computed tomography (CT) scans of the head, electroencephalography, evoked potentials, extra- and intracranial vascular ultrasound, ECG, a negative nitroglycerin and indomethacin test, a psychiatric evaluation, neuropsychological testing, neuroautonomic assessment, routine laboratory blood tests and a spinal tap, all without clear pathological findings.

In 2008, he was diagnosed with having trigeminal neuralgia in another centre, as an MRI had suggested a possible contact between the superior cerebellar artery and the trigeminal nerve. Because of the lack of a clinical improvement under medication, he was offered microvascular decompression surgery of the trigeminal nerve root entry zone (Janetta operation), which was performed by a neurosurgeon with long-standing experience in this procedure. Intraoperatively, the contact of the superior cerebellar artery and the trigeminal nerve was confirmed as well as a perforation of the nerve by a vein. Subsequently, the nerve was detached and the vein was coagulated without complications. The postoperative course was unremarkable. The headache improved for two weeks, but then resumed to the former level and phenotype.

Daily analgesic medication at the time before deep brain stimulation (DBS) surgery included topiramate 150 mg, baclofen 25 mg, phenytoin 300 mg, lamotrigine 1000 mg, aspirin 100 mg, amlodipine 10 mg and carbamazepine 200 mg. On neurological examination, there was a mild postural tremor of the upper limbs and a slow, unstable gait. The neurological examination was otherwise normal including facial cutaneous sensation to pinprick, temperature and light touch; corneal reflexes were normal. The patient fulfilled the International Headache Society (IHS) criteria for SUNCT (ICHD-II code 3.3, ICD 10 G44.08) and the published criteria for DBS surgery in cluster patients (1). Written informed consent on DBS was obtained. Surgery was indicated as an individual (orphan) treatment attempt that did not require a vote of the ethical committee according to German legislation.

Neurosurgical procedure

Because of an implanted cardiac pacemaker, the procedure had to be changed from the previously described procedure for hypothalamic DBS, as a complete preoperative MRI could not be performed (2). The stereotactical planning had to be performed using a T1-weighted sequence. Also, owing to brain atrophy, published coordinates could not be used (3,2,4). It was therefore necessary to use indirect landmarks and compare their localization with the Schaltenbrand-Wahren Atlas to define the target point (5). Because of this procedure the following coordinates in relation to mid-AC-PC-point were used: x = 5 mm, y = 0 mm, z = −5 mm.

Macrostimulation was performed on various different levels from 2 mm above to target level to check for intrinsic effects and side-effects. At ± 1.5 mm in front of the target point, stimulation reproducibly led to an increase in the patient’s alertness and vigilance (4 V), upper lateroversion of the eyes to the left and an increase in blood pressure (5.5 V) as well as the feeling of fear and panic (6.5 V). Microelectrode recordings at the target level revealed single unit activity with a tonic discharge. Intraoperative testing using sensory, motor and emotional stimulation did not reveal obvious neuronal responses that could be attributed to the afferent stimulation. Postoperative imaging confirmed the electrode location in the target area corresponding to the planned coordinates (Figure 1). OPEN IN VIEWER

Clinical effects of neurostimulation

The patient reported a considerable decrease in his attack frequency after initiating hypothalamic stimulation (Figure 1). In the follow-up phase up to 15 months after DBS the attacks remained at a low frequency. However, despite successive increases in the stimulation parameters, it was not possible to reduce the oral medication as changes in dosage led to a subsequent increase of attack frequency. Due to the stimulation effects, he refused to have the stimulator switched off in a patient-blind manner.

During the first 6 months postoperatively the patient regained 12 kg of his body weight. His wife noted a stabilized gait and a balanced mood. The stimulation was well tolerated. Only transient and mild side-effects were noted. The most frequent reported side-effect when initiating neuromodulation or reprogramming was short lasting vertigo. The current stimulation settings are 3 V, pulse width 60 µs, 130 Hz (0−/2+).

Discussion

DBS of the posterior hypothalamic area is an invasive therapeutic approach in patients with intractable chronic cluster headache, and long term results have recently been reviewed (6). However, not all patients respond to DBS (7). Similarly, a beneficial effect of DBS of the posterior hypothalamic area has also been described in two patients with medically refractory and drug-resistant SUNCT using the same stereotactical coordinates as in cluster headache (3,8). A beneficial effect of posterior hypothalamic stimulation has also been described in a patient with chronic paroxysmal hemicrania (9). SUNCT, paroxysmal hemicrania and cluster headache belong to the family of trigeminal-autonomic cephalgias, and in these disorders an activation of the posterior inferior hypothalamus was observed in functional imaging (10). Hypothalamic DBS was first shown by Leone et al. to have an effect on the frequency of headache attacks using the coordinates of hypothalamic activity in PET as target region for DBS in cluster headache and SUNCT (11). However, there are also possible complications with this procedure, as Schoenen et al. reported one fatal intracerebral haemorrhage among a series of cluster headache patients (12). Recently, the anatomical locations of the target electrodes in hypothalamic DBS has been revisited in detail, showing that location of DBS electrodes implanted were not significantly different between responders and non-responders and that electrodes were placed not directly in the hypothalamus but posterior to the mammillary bodies, at the diencephalo–mesencephalic junction (13). Thus, it has been suggested that the DBS in trigeminal-autonomic cephalgias is rather a ‘DBS of the posterior hypothalamic region’, acknowledging the complex nomenclature and anatomy of this region (14).

One of the main differential diagnoses of SUNCT is trigeminal neuralgia, and finding the diagnosis can be challenging in some cases as there is some overlap of symptoms (15). Similar to our case, Black and Dodick reported no beneficial effect of microvascular decompression surgery in SUNCT thus supporting, in conjunction with functional imaging results and effects of DBS, a central nervous system aetiology that can be, however, modulated by peripheral trigger (16,17).Nevertheless, DBS in trigeminal-autonomic cephalgias should still be considered an experimental therapeutic option and should be restricted to patients with a chronic form of this disorder. Careful data collection in an international data bank might be helpful to further elucidate aetiology and treatment options in these rare cases.