Microvascular decompression of the pterygopalatine ganglion in patients with refractory cluster headache

Abstract

Background: Cluster headache is an invalidating form of headache. Although cluster headache can be managed pharmacologically, some patients require surgical treatment with varying results. Microvascular decompression of the pterygopalatine ganglion could be an alternative to traditional surgical management in patients with cluster headache.

Methods: Microvascular decompression of the pterygopalatine ganglion was performed in three patients with refractory cluster headache. The pterygopalatine artery was ligated and a temporal muscle graft was placed between the artery and the ganglion.

Results: No differences were found between the presurgical period and 1 week, 1 month, 3 months, and 6 months postoperatively with respect to attack duration and frequency, visual analogue scale score during attacks and in remission periods, duration of remissions, and quality of life.

Conclusion: These preliminary data suggest that microvascular decompression of the pterygopalatine ganglion does not provide pain reduction or improvement of quality of life in patients with refractory cluster headache.

Introduction

Cluster headache (CH) is a disabling form of headache, characterized by episodic attacks of unilateral pain in association with ipsilateral cranial autonomic features. The specific cause of CH remains unknown. Besides a dysfunction in the central nervous system, the pterygopalatine ganglion (PPG) plays an important role in its pathophysiology (1).

Although CH can be managed pharmacologically, some patients require surgical treatment because of intractable symptoms or adverse effects of medication. Some of these patients may be treated successfully with surgical interventions directed to the PPG. Previous studies have shown a modest effect of radiofrequency thermocoagulation (RFT) of the PPG with success rates ranging from 61 to 65%, but pain relief is often temporary and repeated procedures are required (2,3). Clearly, there is a need for alternative surgical procedures in management of CH, providing longer lasting pain relief.

Examination of the anatomic relations of the PPG reveals that it lies in close contact with a remarkably tortuous portion of the pterygopalatine artery along its course in the PPF, suggesting vascular compression of the PPG as a causative factor in CH (4). Vascular compression is associated with other disorders such as trigeminal neuralgia, hemifacial spasm, and glossopharyngeal neuralgia (5,6). These syndromes can be treated successfully with microvascular decompression (MVD), i.e. discontinuation of contact between the vascular loop and neural structure of interest (7–9). Possibly, discontinuation of contact between the pterygopalatine artery and the PPG could provide pain relief in patients with CH. Here, we report MVD of the PPG in three patients with refractory CH.

Methods

Patients

We conducted a pilot study between October 2008 and May 2010 in which three patients underwent MVD of the PPG. The pilot study was approved by the Medical Research Ethics Committe of our centre, the University Medical Centre Utrecht, the Netherlands. Patients from the department of Otolaryngology of our centre, suffering from chronic CH refractory to extensive pharmacological treatment by neurologists elsewhere, were eligible. Initially, we planned to enrol a total of six patients. Exclusion criteria were pain attributable to other diagnoses, major anatomical variations in the area of interest, disease of the maxillary sinus, previous surgery in the area of interest, and contraindications for MRI.

Patient 1 was a 57-year-old woman with refractory, mostly right-sided CH. She had been diagnosed with chronic CH 5 years earlier by a neurologist elsewhere and had been treated with various medications without adequate pain relief. Cluster attacks were closely spaced, with a frequency of eight a day, a duration of 3 hours, and remission periods of 14 days. Mean score of intensity of pain during attacks on a Visual Analogue Scale (VAS) was 8.

Patient 2 was a 50-year-old woman with refractory, mostly right-sided CH. She had been diagnosed with chronic CH 12 years earlier by a neurologist elsewhere, and was currently treated with imigran injections, without adequate pain relief. Attacks were closely spaced, with a frequency of 10 a day, a duration of 1 hour, and remission periods of a few days. Mean VAS score during attacks was 8.

Patient 3 was a 33-year-old male with refractory, mostly left-sided CH. He had been diagnosed with chronic CH 14 years earlier by a neurologist elsewhere, and had been treated with various medications and two RFT procedures of the PPG 5 years earlier, without adequate pain relief. He had experienced some partial and transient pain relief after the first RFT procedure, lasting about 2 months, but no pain relief after the second procedure. From then on, he had been suffering from medically intractable CH. We decided to enrol patient 3 because MVD is a technique different from RFT, with a different specific aim, i.e. discontinuation of contact between a vascular structure and the PPG, instead of lesioning the PPG itself. Attacks were closely spaced, with a frequency of 2–5 a day and a duration of 2–3 hours and remission periods of 7–14 days. Mean VAS score during attacks was 9.

Procedure

Magnetic resonance imaging (MRI) of the PPF was performed preoperatively in order to study its anatomic relations and possible variations and to exclude concurrent causative pathology. MVD was performed unilaterally on the main side of CH attacks, under general anaesthesia.

After mucosal incision in the buccogingival sulcus, mucosa and periosteum were elevated off the anterior maxillary sinus wall. The infraorbital nerve was identified to avoid its injury, and a 2–3 cm antrostomy was created. The posterior maxillary sinus wall was identified, and a hatch was created in its centre approximately 1–2 cm below the orbital floor. The PPF was exposed and the pterygopalatine artery was ligated with haemoclips (Figure 1). Next, a temporal muscle graft was placed between the artery and the PPG (Figure 2).

Figure 1.

Ligation of the pterygopalatine artery with haemoclips. Arrows point toward haemoclips.

Figure 2.

Temporal muscle graft, placed between the ptergygopalatine artery and the pterygopalatine ganglion. PA, pterygopalatine artery; TMG, temporal muscle graft.

Outcome

Outcome measures were pain and quality of life. Pain was assessed with respect to intensity, frequency and duration. At inclusion, patients filled out a disease-specific questionnaire including information on attack duration, attack frequency, mean VAS score during attacks and in remission periods, and duration of remissions. From the first postoperative day, patients kept a diary of pain intensity assessed on a VAS, during attacks and in remission periods, presence of attacks, and their duration.

Quality of life was based on a generic health-related quality-of-life questionnaire, SF-36, which was filled out at inclusion and at 1 week, 3 months, and 6 months postoperatively. Patients visited the outpatients department for evaluation at 1 week, 6 weeks, 3 months, and 6 months follow up.

Results

None of the patients showed abnormalities of the PPF on MRI. All patients underwent MVD of the PPG: patient 1 and 2 underwent right-sided MVD, patient 3 underwent left-sided MVD. Procedure 1 was complicated by postoperative haemorrhage of the wound bed, treated by electrocoagulation under general anaesthesia. Procedure 2 and 3 were uncomplicated. All patients could be discharged from our clinic on the third postoperative day.

In all three patients, no differences were found between the presurgical period and 1 week, 1 month, 3 months, and 6 months postoperatively with respect to attack duration, attack frequency, mean VAS score during attacks and in remission periods, and duration of remissions. Quality of life did not change between the presurgical and postoperative follow-up period.

In the first month, patient 1 even experienced an increase in attack frequency. Fortunately, the attacks had decreased to their usual frequency by the third month. Patient 1 also reported a temporary hypaesthesia of the infraorbital nerve area, which had resolved at 6 months postoperatively. Patient 2 experienced a slight decrease in attack frequency in the first postoperative week. Unfortunately, after the first week the attacks increased to their usual frequency.

Although we did continue follow-up of the treated patients up to 6 months, according to our pilot study design, we decided to discontinue enrolment of any other patients within this period because of disappointing initial results.

Discussion

This report describes three cases of chronic refractory CH, despite treatment with various suitable medications and, in one case, surgical treatment with RFT of the PPG. No differences were found between the presurgical and postoperative follow-up period with respect to attack duration, frequency, pain, and quality of life scores. Because of disappointing initial results, we decided to discontinue enrolment of other patients within a few months after the first three procedures.

About 90% of CH patients suffer from the episodic form, which can be managed pharmacologically with a variety of medications. However, an unfortunate 10% suffer from chronic CH, in which the attacks are closely spaced with no periods of remission lasting longer than 14 days or pain without remission for greater than a year. Once the chronic CH is established, medical treatment is of limited success (10).

Surgical treatment for refractory CH has remained a frustrating endeavour. Treatment directed to the PPG such as RFT, phenolization, or direct ganglioneurectomy, as well as section of the intermediate nerve (11) or superficial petrosal nerve, have been performed previously without providing long-lasting pain relief (2,3).

As MVD of the PPG has not been described in management of CH before, our findings cannot correspond to those in previous studies. In trigeminal neuralgia, previous studies have shown that MVD of the trigeminal nerve provides longer lasting pain relief when compared to RFT of the PPG (7,8). Previously, Lovely et al. (10) reported successful surgical management of chronic CH through MVD of the trigeminal nerve, alone or in combination with section and/or MVD of the intermediate nerve in 28 patients, including two with bilateral disease. However, although 22 of 30 procedures (77.3%) resulted in an excellent or good outcome in the immediate postoperative period, the success rate dropped to 46.6% with long-term follow up, i.e. after the first postoperative year.

The negative results of our study could be due to technical surgical aspects. More likely, however, our hypothesis of vascular compression of the PPG as a causative factor in CH, is not as strong as assumed. Probably, other factors such as dysfunction in the central nervous system are of greater importance in the pathogenesis of CH.

A limitation of our study is that the number of participants is extremely small. However, due to disappointing initial results, enrolment of more patients did not seem acceptable. Furthermore, the study was designed as a pilot with a maximum of six patients, to be continued if initial results would be promising.

This report is the first of its kind describing MVD of the PPG in the surgical management of refractory CH. Unfortunately, our preliminary data suggest that MVD of the PPG does not provide pain reduction or improvement of quality of life in patients with refractory CH.