Paroxysmal sneezing after hypothalamic deep brain stimulation for cluster headache

Abstract

Background: Cluster headache (CH) is the most common of the trigeminal autonomic cephalalgias (TAC), presenting with excruciatingly severe, short-lasting, unilateral headache accompanied by cranial autonomic symptoms. Chronic CH occurs in 10–15% of patients. Deep brain stimulation in the posterior hypothalamic region (hDBS) is successful in treating about 60% of patients otherwise refractory to medical treatment.

Case: A 28-year-old man had hDBS for medically refractory left-sided chronic CH, with a resultant reduction in frequency and severity of his attacks. He developed recurrent paroxysms of sneezing soon after the stimulation was started that have reduced after increasing the pulse width from 60 to 90 µs.

Discussion: Stimulation of the brain in the region of the posterior hypothalamus could produce sneezing from activation of facial nerve parasympathetic or trigeminal afferent pathway activation through the trigeminohypothalamic tract, or through other central mechanisms. DBS in general offers the opportunity to illuminate our understanding of brain function and for CH offers particular opportunities to understand a devastating primary headache syndrome.

Background

Cluster headache (CH) is the most common trigeminal autonomic cephalalgia (TAC) with attacks of short-lasting, unilateral, extremely severe pain with cranial autonomic symptoms. The International Classification of Headache Disorders (ICHD-2) divides CH into episodic and chronic subtypes (1). The diagnosis of chronic CH requires pain-free periods of less than 1 month in the previous year and affects about 10–15% of sufferers. Over the last decade neuromodulatory approaches to the treatment of medically refractory CH have been developed, particularly deep brain stimulation in the posterior hypothalamic region (hDBS) (2), and occipital nerve stimulation (3). We report a case of persistent sneezing developing after hDBS for chronic CH. As far as we are aware, this has not been previously reported.

Case

A 28-year-old man had DBS in the posterior hypothalamic region in July 2009 for treatment of refractory left-sided chronic CH. The anatomic target for DBS on the left side was: 2 mm posterior, 5 mm inferior and 2 mm lateral to the mid-commissural point. The pain was side-locked over the left supraciliary region, with interictal constant pain in the same area. The interictal pain started 2 years after the episodic left supraciliary headache, was rated as moderate in intensity and there were no accompanying migrainous features. The problem had started 5 years previously with significant worsening 2 years prior to the DBS. His other medical problems included restless legs syndrome (which his father also had), hypertension, depression, and obstructive sleep apnea. He had a strong family history of migraine. All four of his children have eosinophilic enteropathy; he tested negative for this on upper and lower GI endoscopy and biopsies. Oxygen 100% at 15 L/minute, sumatriptan nasal spray and 6 mg injections, zolmitriptan nasal spray, dihydroergotamine nasal spray, fentanyl patch 400 µg, oral oxymorphone 30 mg TDS, methadone 30 mg OD and oral hydromorphone 4 mg had failed to abort the pain. Meperidine injections reduced the intensity of the pain to some extent. For prophylaxis, he was treated unsuccessfully with topiramate 200 mg twice daily for 2 months, verapamil 300 mg twice daily for 1 year, prednisolone 100 mg tapered off over 10 days, methergine 0.2 mg three times daily, propranolol 50 mg twice daily for 2 months, lamotrigine 75 mg daily (stopped due to rash), gabapentin, valproate, pregabalin, indomethacin for 3 months (making hemicrania continua unlikely) as well as dronabinol 2.5 mg three times daily for nausea. Although lithium 900 mg twice daily for 3 months improved the intensity of his headaches by a small margin without improving the frequency, he developed mental slowness and tremor and it was discontinued. A greater occipital nerve block with lidocaine and corticosteroid, an inpatient admission for intravenous dihydroergotamine and a stellate ganglion block did not help.

The pain improved 3 weeks after the first DBS in July 2009 but unfortunately he developed an MRSA infection of the extra-cranial part of the electrode in the frontal region and Stevens–Johnson syndrome as a reaction to antibiotics. The lead was removed in August 2009. A repeat procedure was done in March 2010 and the lead was re-sited with the anatomic target: 2 mm posterior, 5 mm inferior and 2 mm lateral to the mid-commissural point (Figure 1). This was also 4 mm posterior to the left mamillothalamic tract. The initial settings were voltage 2.0 V, contact 1 monopolar, frequency 185 Hz and pulse width 60 µs. The electrode tip appeared to be in the same location in the post-operative MRI scans on both occasions, although it is possible that there was a minute variation in the position. When the DBS was turned on 10 days after insertion, he started sneezing within hours. He had repeated paroxysms of sneezing every couple of hours, 7 or 8 times a day in total. During each paroxysm, he sneezed several times and profusely, which was quite disconcerting for him. The sneezing did not occur after the first DBS. The patient switched off the stimulator for few hours but this did not make a difference to the sneezing. Within a month he had some relief in the background interictal pain as well as the intensity and duration of the episodic attacks. On a follow-up visit a couple of months after insertion, the amplitude was increased to 2.4 V. This brought about a significant improvement, mainly in the frequency of the CH attacks, reducing from 8–10 a day to 3–6 a day. He continued to have 7–8 severe paroxysms of sneezing a day until December 2010, when the pulse width was increased from 60 to 90 µs. This has reduced the frequency of sneezing to 1–2 times a day and they are not as severe (Table 1).

Figure 1.

Post-operative T2 scan showing lead placement in left hypothalamic region (as indicated by arrow).

Table 1.

hDBS parameters – effects on sneezing

hDBS on/off	Voltage (V)	Frequency (Hz)	Contact	Pulse width (µs)	Headache status	Sneezing
First hDBS – on	1–1.5	185	Monopolar	60	Frequency of episodes reduced from 8 to 3–4/day and intensity from severe to moderate, after 3 weeks; interictal pain unchanged	None
Between first and second hDBS					Interictal pain present, frequency of episodes 8–10/day, severe pain	None
Second hDBS – off for 10 days after procedure	–	–	–	–	Unchanged from above	None
Second hDBS – on	2.0	185	Monopolar	60	Interictal pain, and intensity and duration of episodes reduced after 4 weeks	Yes – recurrent paroxysms, 7–8/day
Second hDBS – off for few hours	–	–	–	–	Unchanged from above	Unchanged from above
Second hDBS – on	2.4	185	Monopolar	60	Frequency of episodes reduced from 8–10 to 3–6/day	Unchanged from above
Second hDBS – on	2.4	185	Monopolar	90	Unchanged from above	Reduced to 1–2/day

Discussion

DBS for chronic CH is generally considered safe and effective in about 60% of patients, as shown by a recent review of 50 cases (2), although one patient died from intracerebral hemorrhage several hours after the procedure (4). Short-lasting diplopia is the most common side effect, limiting increase in the strength of hypothalamic stimulation (4). No changes in body temperature, electrolyte balance, sleep–wake cycle, blood pressure, appetite or thirst have been noted after prolonged stimulation (5). Pituitary hormones, testosterone and EEG are not altered (5). Continuous unipolar stimulation of the hypothalamus has been noted to increase sympathetic-excitatory effects during HUTT (tilt-table) but Valsalva maneuver, deep breathing, cold-face test, isometric hand-grip and baroreflex sensitivity were not affected (5). A single case of bradycardia during the procedure was noted (5). The literature does not suggest any effect on respiratory function or sneezing.

Sneezing is a semi-autonomous act of forcefully exhaling air through the nose and mouth to clear the airway passages of irritants. The afferent pathway is thought to be branches of the trigeminal nerve supplying the nasal mucosa. In animal studies these nerve fibers relay to the ventromedial spinal trigeminal tract and the adjacent medullary reticular formation (6). In humans the lateral medulla is important in sneezing. Sneezing is absent in patients with lateral medullary syndrome (7) and medullary neoplasm (8) and paroxysmal sneezing was noted in a patient with a lateral medullary infarct (9). The efferent pathway is the pharyngeal, laryngeal and chest wall muscles, which forcefully contract to expel air at speeds of around 150 km/hour, with the soft palate and uvula depressed to expel most of the air through the nose. Parasympathetic nervous system mediated increases in nasal and lacrimal secretions commonly occur during sneezing.

We illustrate the following conditions where parasympathetic stimulation can lead to sneezing. Sneezing induced by photic stimulation is well known and can be inherited as an autosomal-dominant condition (10). Sneezing occurs after heavy meals in some individuals, which also has an inherited aspect (11). Sneezing has been reported as induced by sexual ideation or orgasm; this may be under-reported (12). One of the common things in these three conditions associated with sneezing is parasympathetic activation. Exposure to bright light causes meiosis, a full stomach stimulates peristalsis and acid secretion, and sexual ideation leads to venous dilatation leading to penile and clitoral tumescence; all of these phenomena are mediated through parasympathetic activation. The vagal center in the medulla plays a central role in parasympathetic innervation and it is possible stimulation in one area may inadvertently cause stimulation in another area. Thus, nasal mucosal irritation mediated by the facial parasympathetic innervation can occur in these conditions, which can stimulate the afferent arc of the sneezing reflex.

The hypothalamus is the regulatory center for the autonomic system. The anterior part is thought to regulate the parasympathetic and the posterior part of the sympathetic system. However, this distinction may not be completely correct. Animal experiments have shown projections from anterior and posterior hypothalamus to the dorsal vagal nucleus in the medulla (13). In our patient placement of the lead in the posterior hypothalamic region may have caused parasympathetic stimulation leading to nasal mucosal irritation via the facial parasympathetic outflow with activation of the afferent arc of the sneeze reflex via the trigeminal nerve. It must be said that autonomic testing after hDBS for CH has not revealed any significant parasympathetic effects, so this may be particular to our patient or to the electrode placement. Another possible mechanism is direct hypothalamic influence on trigeminal nerve function. There is animal experimental evidence of the presence of a trigeminohypothalamic tract (14). Moreover, human studies showed an increase in cold pain threshold in the distribution of the first division of the trigeminal nerve on the same side after hDBS for chronic CH as compared to non-implanted CH patients and healthy controls (15).

The patient switched off the stimulator for few hours but this made no difference to the sneezing. Effects of neurostimulation are known to last beyond the period of stimulation and this may be the reason for above. The other possible explanation is that of tissue damage by the electrode rather than stimulation itself. Although the electrode tip is thought to be in the posterior hypothalamic region, it is possible that the tip is in the ventral tegmental region of midbrain, posterior to the posterior boundary of the hypothalamus. This region is not directly known to influence sneezing. However, the effects of stimulation involve a wider area of the brain, which may possibly involve sneezing. For example, the periaqueductal gray (PAG) is known to have a variable influence on respiratory patterns, including sneezing, possibly mediated through the nucleus retroambiguus in the medulla (16). A recent case report points to the improvement of pre-existing polydipsia and restlessness but not of CH after hDBS, which also caused a high-frequency tremor suggestive of effects on functions other than pain (17). Lastly, in relation to the improvement in sneezing after increasing the pulse width and not the amplitude of stimulation, there may be subtle differences between changes in pulse width vs. amplitude, with respect to the exact neural elements excited, but both maneuvers increase current, and in general, increase the volume of tissue activated.

The case illustrates that there is much to be learnt from the treatment of primary headache syndromes where brain neuromodulation techniques are employed. Such cases offer a unique ability to understand better brain function and the influences of various central nervous system structures and their role.