Keywords

10.1 Introduction

For more than 100 years, the sphenopalatine ganglion (SPG) has been targeted for the treatment of headache and facial pain. Several techniques have been used over the years to influence the activity in this neuronal structure, from intranasal cocaine to today’s implanted stimulators. Data collected throughout the years point in the direction of an important role of this parasympathetic ganglion in different primary headaches. The aim of the present chapter is to give an overview of the anatomy and physiology of this structure with relation to headache pathophysiology and in particular how it may be targeted to treat headache disorders.

10.2 Anatomy

The SPG is located in the pterygopalatine fossa, a triangular space with a volume between 0.1 and 1 cm3 [1] (Fig. 10.1). In most individuals (70%), the ganglion is a single, 3–4 mm long (in the craniocaudal direction), conical structure, whereas in around 30% it consists of two separate parts [2]. It contains sensory, sympathetic, and parasympathetic nerve fibers. The sensory fibers enter the ganglion as branches of the maxillary nerve, pass through without synapsing, and exit as the lesser and greater palatine nerves, supplying sensation to the palate and the mucous membrane of the nasal septum [3]. The postganglionic sympathetic fibers from the superior cervical ganglion run through the internal carotid plexus to form the deep petrosal nerve, which merges with the parasympathetic greater petrosal nerve to form the Vidian nerve. The sympathetic fibers of the Vidian nerve, passing through the SPG without synapsing, are distributed to the nose, palate, and lacrimal gland [4]. Preganglionic parasympathetic fibers originating in the superior salivatory nucleus (SSN) in the pons run through the facial nerve and further to the geniculate ganglion from which they exit in the greater petrosal nerve. This nerve merges with the deep petrosal nerve to form the Vidian nerve that enters the SPG where they synapse. The postganglionic fibers coming from the ganglion are distributed to the nose, palate, and lacrimal gland, as well as to cerebral and meningeal vessels [2].

Fig. 10.1
figure 1

Anatomy of the sphenopalatine ganglion (Gray’s Anatomy for Students, 2nd Edition. 2009, ELSEVIER)

Full size image

Recent studies have demonstrated that there are small calcitonin gene-related peptide (CGRP) containing neurons in the SPG that most likely originate from the trigeminal ganglion [5], indicating that there may be a direct interaction between the sensory system and the SPG [6].

10.3 Physiology

The trigeminal-parasympathetic reflex center in the brainstem connects trigeminal afferents and the parasympathetic efferents that synapse in the SPG [7, 8]. Activation of the parasympathetic efferent neurons leads to perivascular neurotransmitter release (vasoactive intestinal peptide (VIP), nitric oxide, and acetylcholine), resulting in dilatation of intracerebral blood vessels, plasma protein extravasation, and activation of meningeal trigeminal nociceptors [9, 10]. The current rationale for targeting the SPG in headache conditions is to interrupt parasympathetic outflow and thereby inhibit activation of trigeminal afferents.

10.3.1 Why Are We Targeting the SPG?

Several techniques for disrupting neuronal signaling in the SPG have been used over the years, including application of chemical substances (cocaine, alcohol, local anesthetics, steroids, botulinum toxin) in or near the ganglion, nerve sectioning, radiofrequency ablations, and electrical stimulation of the ganglion. Two case reports on stereotactic neurosurgery on the SPG have also been published [11, 12]. Interruption of parasympathetic outflow is considered the most plausible mechanism of action. Today, one used technique for interrupting neuronal activity in the SPG is high-frequency electrical stimulation. This works possibly by depletion of stored neurotransmitters in parasympathetic efferents, leading to reduced activation of meningeal nociceptors [13, 14]. Other modes of action could include an antidromic inhibitory effect on the SSN or even some degree of nerve conduction block. The latter is considered less likely given the stimulation frequencies used today [15]. Another possibility could be sensory modulation through the sensory (maxillary) fibers in the pterygopalatine fossa, as most techniques targeting the SPG are nonselective as to which types of nerve fibers are influenced (Sect. 10.3.2) [8]. The role of the direct neuronal pathways between the parasympathetic and the sensory system (Sect. 10.2) for the clinical effects is not known but should be further explored.

10.3.2 What in the SPG Are We Targeting?

The SPG is traversed by sympathetic, parasympathetic, and sensory fibers. How do we know which fibers are influenced by current techniques used to target the SPG, and which fibers are most important to target in different headache conditions? These important questions have previously been asked by Narouze [16]. Possibly, it could be more effective to selectively influence one specific type of nerve fiber. Which one, could vary according to the condition we are treating. As an example, targeting of the sensory fibers from the second division of the trigeminal nerve might be more relevant for certain types of orofacial pain than for cluster headache or migraine. In addition to type of fiber to target, it is also important to know how the target area responds to the intervention. For instance, what would be the optimal frequency (or dose for chemical agents) for attaining an optimal effect? In a recently published study, low-frequency stimulation of the SPG did not activate Aβ fibers or C fibers when testing for mechanical perception and pain thresholds [17]. High-frequency stimulation, on the other hand, readily elicits sensory symptoms [18]. These data emphasizes the need to determine the thresholds needed (for all types of SPG targeting modalities) to facilitate or inhibit activity in sympathetic, parasympathetic, or sensory fibers.

A main question has been whether parasympathetic activation is involved in the generation or maintenance of pain in different headache conditions, in addition to producing autonomic symptoms. Yarnitsky et al. observed that migraine patients with no parasympathetic symptoms were less likely to experience pain relief after treatment with nasal lidocaine than those with such symptoms [10]. The interpretation was that the parasympathetic pathways could contribute to the pain as well. In cluster headache, further support for the potential importance of the parasympathetic pathways was illustrated by a patient that continued having cluster headache attacks, even after the ipsilateral trigeminal sensory root was excised [19]. Schytz et al. demonstrated that low-frequency (LF) stimulation of the SPG could provoke cluster headache attacks [20]. A recently published study, however, challenges these findings. When the Schytz study was replicated with slightly higher stimulation frequencies and longer duration, LF stimulation did not provoke cluster-like headache significantly more often than sham (35% vs. 25%) [17], although autonomic symptoms did appear more frequently with LF stimulation (80% vs. 45%, p = 0.046). The role of the parasympathetic efferents in pain generation or maintenance is still unclear.

10.4 Cluster Headache

10.4.1 Intranasal Administration of Topical Agents in Cluster Headache

The first description of intranasal administration of a topical agent (cocaine) toward the SPG (n = 5) was made by Sluder in 1908 in “Sluder’s neuralgia” (resembling today’s cluster headache diagnosis) [21]. The apparent effect of cocaine was seen in other similar cases but had to be applied too often, with risk of negative side effects [22, 23]. The nonaddictive alternative, lidocaine, was found to be just as effective in a couple of case series [24, 25]. Further, 88% phenol applied with cotton swabs intranasally also had a relevant effect over time [26]. However, only one controlled study has been performed (n = 15) with intranasal topical agents in cluster headache [27]. Both cocaine and lidocaine showed superiority to saline; with complete cessation of pain after a little more than half an hour for active substances, compared to 1 h for placebo.

10.4.2 Invasive Techniques Targeting the SPG in Cluster Headache

10.4.2.1 Destructive Techniques

Meyer et al. removed the SPG (histologically verified) in 13 patients with cluster headache [28] but with no or only modest clinical effect: seven patients had no effect, four had incomplete relief, and only two had complete relief over the next 12 months. One patient had previously undergone a trigeminal rhizotomy with anesthesia of the maxillary region but no effect on the pain attacks. With the additional SPG resection, there was an initial remission, but later the pain recurred. This demonstrated that even a combined destruction of both the sensory afferent (trigeminal pathways) and the parasympathetic efferent (SPG) was unhelpful. This was in line with previous findings in “sphenopalatine neuralgia” (cluster headache-like symptoms), where stimulation of the greater petrosal nerve provoked pain and sectioning of the nerve in 13 patients provided relief. However, the results obtained were quite variable (“excellent” in 25%, “good to fair” in 50%, and “poor” in 25%) [29].

Later, several other types of irreversible SPG blockades were performed, including supra-zygomatic alcohol injection, relieving pain in 85% of 120 patients with a follow-up time of 6 months to 4 years [30]. However, a disadvantage with this treatment is that the alcohol can spread to the maxillary nerve and cause painful neuritis [31]. The use of radiofrequency techniques has been effective in 60–80% of episodic cluster headache patients and 30–70% of chronic cluster headache patients, in a number of small, uncontrolled studies [31,32,33,34,35]. In order to avoid the more cumbersome fluoroscopy technique, a Chinese group explored the use of CT-guided pulsed radiofrequency treatment in refractory episodic (n = 13) and chronic (n = 3) cluster headache [34]. Eleven of the former and one of the latter patients showed complete remission within 6 days.

10.4.2.2 Neuromodulatory Techniques

With the use of an endoscopic technique and injecting through the nasal wall, an Italian research group deposited a mixture of local anesthetics and corticosteroids toward the sphenopalatine fossa in 20 patients with refractory chronic cluster headache [36]. The treatment resulted in improvement in 58% of the patients, but the effect was short-lasting, and repeated injections were less efficacious. The same technique was further evaluated in 15 patients with chronic cluster headache, where 60% responded to the treatment [37]. In the latter study, there was a tendency to have a more long-lasting response, possibly owing to the improved injection technique.

One of the problems with using local anesthetics has been their short-lasting effect. Aiming at exploiting the longer-lasting effect of botulinum toxin, a small open-label pilot study on ten patients with refractory chronic cluster headache investigated the efficacy and safety of injecting 25–50 units of botulinum toxin toward the SPG on the symptomatic side [38]. The injection was performed with a new surgical device using an image-guided navigation (MultiGuide®). Five of ten patients responded to the treatment with an average attack frequency reduction (main efficacy outcome) of 77% during the 6 months of follow-up period. The safety profile was considered acceptable by the authors. A physiological basis for the use of botulinum toxin toward the SPG was strengthened when the botulinum toxin receptor, synaptic vesicle glycoprotein 2-A (SV2-A), and the synaptosomal-associated protein 25 (SNAP-25) were found in human SPGs [39]. A randomized controlled trial is currently being planned to confirm the results of this preliminary open-label pilot study.

In 2010 six patients with refractory chronic cluster headache received short-term electrical stimulation of the SPG [15]. The acute treatment response was promising. Later, a multicenter, sham-controlled study using an implantable, on-demand SPG stimulator (Pulsante®) in medically refractory chronic cluster headache (Pathway CH-1 study) demonstrated a clinically significant improvement in 68% of the patients. Among these, 25% were acute responders (i.e., had pain relief at 15 min in >50% of treated attacks), 36% were frequency responders (>50% reduction in attack frequency), and 7% were both acute and frequency responders [13]. The observation that there might be a prophylactic effect of the acute treatment was unexpected, but follow-up data support this finding. An open-label follow-up study at 24 months (n = 33) demonstrated a long-term clinical efficacy with 45% acute responders, 33% frequency responders, and 61% either acute responders or frequency responders or both [40]. A recently published open-label study, using the same type of SPG stimulator in 97 cluster headache patients (88 chronic and 9 episodic), investigated efficacy at 12 months [41]. Twelve patients had their stimulators removed due to lack of effect or adverse effects. Of the remaining 85 patients, the mean attack frequency was reduced from 25.2 to 14.4 attacks per week, and 68% were treatment responders (defined as being acute responder or frequency responder or both). Thirty-two percent of all patients were acute responders, and 55% of the chronic patients were frequency responders. The implant procedure requires unique anatomical and surgical skill and is mainly performed by few selected cranio-maxillofacial surgeons in Europe. Therefore, it is recommended that the implants and the follow-up are concentrated to specialized centers with relevant surgical expertise.

A sham-controlled trial testing the Pulsante therapy on 120 patients with chronic cluster headache patients is currently ongoing in the USA (Pathway CH-2 study NCT02168764—ClinicalTrials.gov accessed October 2017).

10.4.3 Clinical Usefulness of SPG-Targeted Treatments in Cluster Headache

  • The evidence does not support the widespread clinical use of intranasal local anesthetics in cluster headache.

  • Endoscopic injection with local anesthetics and steroids appears to have a relatively short-lasting effect.

  • A number of uncontrolled studies have indicated an effect of radiofrequency treatment in cluster headache, but there may be irreversible side effects, and the fluoroscopy method is somewhat cumbersome for widespread clinical use. The use of CT guidance may be an easier technique, but concern about radiation limits the number of repeated injections.

  • The use of botulinum toxin injections toward the SPG is currently being further investigated. In theory, the use of botulinum toxin could have a long-term effect. No controlled studies have yet been performed.

  • SPG stimulation has increasing evidence for long-term efficacy in chronic cluster headache and is providing interesting new insight into the pathophysiology of cluster headache. Overall, this appears to be a safe therapy with tolerable and transient side effects in most patients. However, confirmation of its role in the treatment of chronic cluster headache awaits the results of the ongoing randomized sham-controlled trial (CH-2). Since the implant procedure is relatively complicated and the eligible number of patients low, patient selection and surgery should be centralized to hospitals with high expertise in headache and dedicated surgeons.

10.5 Migraine

It is firmly established that cranial autonomic symptoms (CAS) indicating activation of the trigemino-parasympathetic reflex can be present in both adult [42,43,44,45] and pediatric [46] migraine patients. Yarnitsky et al. showed that migraine patients with parasympathetic symptoms were more likely to experience pain relief 10–30 min after a lidocaine-induced SPG block than those with no parasympathetic symptoms [10]. This could indicate that targeting the SPG in migraine could affect not only parasympathetic symptoms but also influence the pain [8]. In addition, a pathway involving the SPG parasympathetic efferents whereby common migraine triggers activate meningeal nociceptors has been proposed [47]. Stress, lack of sleep, and olfactory stimulation may activate multiple brain areas (hypothalamic, cortical, limbic) with neural connections to the SSN, resulting in activation of postganglionic parasympathetic fibers in the SPG and subsequent activation of meningeal nociceptors [47, 48].

10.5.1 Intranasal Administration of Local Anesthetics in Migraine

The most common way to target the SPG in migraine has been through intranasal application of topical anesthetics, such as lidocaine or bupivacaine. An uncontrolled study on 23 migraine patients in 1995 indicated an effect following application of lidocaine, with 12 patients responding, most within 5 min [49]. Later, five small controlled studies have been performed yielding mixed results:

  1. 1.

    Positive trials:

    1. (a)

      1996 Lidocaine vs. saline with evaluation of effect after 15 min, n = 81. Relief in 55% of the lidocaine group compared to 21% in the saline group, p = 0.04 [50].

    2. (b)

      1999 Lidocaine vs. saline with evaluation of effect after 15 min, n = 131. Relief in 35.8% of the lidocaine group compared to 7.4% in the saline group, p < 0.001 [51].

    3. (c)

      2015 Bupivacaine vs. saline with evaluation of effect after 15 min, 30 min, and 24 h, n = 41. Relief as early as 15 min with a 20% reduction in numeric rating scale (NRS) scores for the bupivacaine group compared to 7% reduction in the saline group [52]. A follow-up of this study was to continue treating the subjects prophylactically twice per week and compare baseline headache days to number of headache days 1 month posttreatment. A significant decrease was not demonstrated [53].

  2. 2.

    Negative trials:

    1. (a)

      2001 Lidocaine vs. saline with evaluation of relief after 5, 10, 15, 20, and 30 minutes, n = 49. No difference between the groups [54]

    2. (b)

      2017 Lidocaine vs. saline with evaluation of relief after 15 and 30 min, n = 162. No difference between the groups [55]

10.5.2 Invasive Techniques Targeting the SPG in Migraine

A pilot study on 11 patients with refractory migraine (9 with medication overuse and 2 with episodic migraine) tested the efficacy of inducing migraine attacks and then treating them with electrical stimulation toward the SPG through an infrazygomatic electrode [56]. Two patients experienced complete remission of the attacks, and three reported meaningful pain reduction. Five patients had no effect and one was not stimulated.

A randomized controlled trial (NCT01540799) testing an SPG stimulator in 80 patients with predominantly fixed (no side-shift) episodic migraine is currently ongoing (ClinicalTrials.gov accessed October 2017).

In an uncontrolled pilot study (n = 10) using a novel image-guided surgical device, 10 patients with refractory chronic migraine were injected with 25 units botulinum toxin toward the SPG on both sides (total 50 units) [57]. The frequency of moderate or severe headache days in the second month post-injection (primary efficacy measure) was 53% less than at baseline, p = 0.009.

10.5.3 Clinical Usefulness of SPG-Targeted Treatments in Migraine

  • It has not been proven that intranasal topical local anesthetics blocks the SPG parasympathetic, sensory, or sympathetic fibers.

  • Controlled studies on intranasal administration of local anesthetics in migraine report conflicting results. Better data are needed before this method is implemented into routine-clinical practice.

  • There are ongoing studies evaluating the efficacy and safety of SPG stimulation in patients with episodic migraine as well as investigation of bilateral injection of botulinum toxin in patients with chronic migraine.

10.6 Other Pain Conditions

A small placebo-controlled crossover study, attempting to target the SPG maxillary nerve fibers in 25 patients with idiopathic second branch trigeminal neuralgia, was performed with 8% lidocaine nasal spray versus saline [58]. Pain recorded on a visual analogue scale (VAS) was reduced from 8.0 at baseline to 1.9 at 15 min after treatment in the lidocaine group (p < 0.01). For saline, the numbers were 7.9 at baseline versus 7.6 posttreatment. The effect lasted median 4.3 h (range 0.5–24 h).

Noninvasive management of persistent idiopathic facial pain (PIFP) is the preferred treatment in most cases. A retrospective study on the use of alcohol injections toward the SPG in 42 patients with refractory facial pain showed that 67% of the treatments were deemed as successful with a mean pain relief period of 10.3 months [59]. The patients who were classified as PIFP (n = 10) had the highest success rate (85.7% of treatments).

A randomized, placebo-controlled study with intranasal administration of bupivacaine versus saline in 93 patients with acute frontal headache in the emergency department did not detect any difference between active and placebo treatment [60]. The primary endpoint was 50% reduction in pain at 15 min, and this was found in 48.8% of the bupivacaine group and 41.3% of the saline group.

10.7 Cranial Parasympathetic Innervation

The SPG is the main source of parasympathetic innervation to the head. Another cranial parasympathetic ganglion, which has attracted much less attention than the SPG, is the otic ganglion (OTG). There is evidence that postganglionic parasympathetic fibers from the OTG also innervate intracerebral vessels [61,62,63,64]. Could there be anything to gain by blocking the OTG or perhaps even a dual block of the SPG and the OTG? Goadsby et al. showed that approximately 50% of the parasympathetic activity to the cranial vessels in cats was conveyed through the OTG and 50% through the SPG [65]. The exact importance of the OTG in humans is not known, but there are data indicating that parasympathetic fibers from the SPG may be more important for the frontal region of the cranium, whereas such fibers from the OTG are more important for the occipital region [10, 66]. Neurochemical studies have demonstrated co-localization of both pituitary adenylate cyclase-activating polypeptide (PACAP) and vasoactive intestinal peptide (VIP) in both the SPG and the OTG [67,68,69]. Hopefully future studies will shed more light on whether the OTG, which has many similarities with the SPG, has any role in headache pathophysiology and whether it may be a feasible and useful target for therapy.

10.8 Conclusion

The sphenopalatine ganglion appears to have a central role in cluster headache and potentially in other primary headache and idiopathic facial pain disorders. New technology that provides a more accurate technique to target the SPG offers the opportunity to evaluate the efficacy of various treatments that can be more precisely directed toward the SPG. This will also facilitate a better understanding of the role of the SPG in the pathogenesis of a variety of headache and craniofacial pain disorders.