Sage Journals: Discover world-class research

Abstract

Background

Low frequency (LF) stimulation of the sphenopalatine ganglion (SPG) may increase parasympathetic outflow and provoke cluster headache (CH) attacks in CH patients implanted with an SPG neurostimulator.

Methods

In a double-blind randomized sham-controlled crossover study, 20 CH patients received LF or sham stimulation for 30 min on two separate days. We recorded headache characteristics, cephalic autonomic symptoms (CAS), plasma levels of parasympathetic markers such as pituitary adenylate cyclase-activating polypeptide-38 (PACAP38) and vasoactive intestinal peptide (VIP), and mechanical detection and pain thresholds as a marker of sensory modulation.

Results

In the immediate phase (0–60 min), 16 (80%) patients experienced CAS after LF stimulation, while nine patients (45%) reported CAS after sham (p = 0.046). We found no difference in induction of cluster-like attacks between LF stimulation (n = 7) and sham stimulation (n = 5) (p = 0.724). There was no difference in mechanical detection and pain thresholds, and in PACAP and VIP plasma concentrations between LF and sham stimulation (p ≥ 0.162).

Conclusion

LF stimulation of the SPG induced autonomic symptoms, but no CH attacks. These data suggest that increased parasympathetic outflow is not sufficient to induce CH attacks in patients.

Study protocol

ClinicalTrials.gov registration number NCT02510729

Keywords

Cluster headache neurostimulation sphenopalatine ganglion headache model autonomic nervous system

Introduction

The phenotypic hallmark of cluster headache (CH) is severe head pain and prominent autonomic symptoms such as lacrimation, conjunctival injection, eyelid edema, rhinorrhea, and forehead/facial sweating (1,2). These cephalic autonomic symptoms (CAS) suggest increased parasympathetic outflow from the superior salivatory nucleus (SSN) via the sphenopalatine ganglion (SPG) (3).

Targeting the SPG by electrical stimulation has emerged as a novel treatment strategy in patients suffering from CH (4 –6). The rationale behind the efficacy is that high frequency (HF) (120 Hz) stimulation of the SPG may abort a cluster attack by blocking parasympathetic outflow. We reported that low frequency (LF) (5 Hz) stimulation induced cluster attacks in patients with chronic CH (7). However, this study was limited by a small sample size of six patients, who also had a heterogeneous response to SPG stimulation therapy. Since this study, more CH patients have been implanted with the SPG stimulator (Pulsante™ SPG Microstimulator System), allowing us to investigate the attack-inducing effects of LF SPG stimulation in a large sample size.

We hypothesized that LF stimulation of the SPG would increase parasympathetic outflow, activate sensory afferents, and provoke a cluster-like attack. We conducted a double-blind, randomized, sham-controlled, crossover study to investigate cluster attack induction following LF (20 Hz and 0.2–2.1 amplitude) and sham (amplitude = 0) stimulation in 20 CH patients with implanted SPG neurostimulators. In addition, we measured plasma concentrations of pituitary adenylate cyclase-activating polypeptide-38 (PACAP38) and vasoactive intestinal peptide (VIP), as markers for parasympathetic activation before, during, and after SPG stimulation.

Materials and methods

We recruited CH patients implanted with the SPG stimulator from the Pathway CH-1 study (6) and the Pathway R-1 study at the Danish Headache Center. Four patients were from the Pathway CH-1 study and 16 were from the registry study. Patient selection criteria for the CH-1 study are that they should be chronic CH patients who are inadequately treated with available therapies (6). Selection criteria for patients in the R-1 study are that they should meet CH criteria per CE-labeling, with clusters lasting a minimum of 16 weeks. The exclusion criteria were: Changes in preventive medication in the month before enrollment; medication overuse headache; pregnancy/lactation; significant somatic and psychiatric diseases. We enrolled patients irrespective of their headache response to therapeutic stimulation. Therapy settings are individualized and stimulation (LF or sham) can be adjusted by physicians using a programmer laptop. More information about the materials and methods can be found in the supplementary material.

Design

We randomly allocated 21 CH patients with an implanted Pulsante system to receive LF SPG neurostimulation (20 Hz) or sham stimulation (amplitude = 0) for 30 min on two experimental days separated by at least 5 days. The randomization was balanced, in which an equal number of patients received LF or sham stimulation on the first experimental day. The patients were not told that one of the stimulations was sham, but only that we would stimulate using frequencies that were different from their usual therapeutic HF stimulation.

All patients arrived non-fasting at the clinic between 8:30 a.m. and 5:00 p.m. CH attacks occurring less than 3 h before the start of the study resulted in postponement of the experiment. Use of oxygen, triptans or neurostimulation 3 h prior the start of the study was not allowed, but the patients could use prophylactic headache medication such as verapamil. Pulse width and electrode configuration were kept as the usual settings of each patient. An unblinded investigator (SG) programmed the stimulator according to a balanced randomization code generated by online software (www.randomizer.org), whereas the patients and investigator (ASP) recording the data were blinded to the neurostimulator settings.

Patients were kept in the supine position in a quiet room with a constant temperature and after 15 min of rest, baseline recordings were performed. Immediately after baseline recordings, LF SPG stimulation (20 Hz) or sham (amplitude = 0) was applied for 30 min. Headache characteristics, CAS and vital variables were recorded at baseline and then every 10 min until 1 hour after start of stimulation. Von Frey hair and pinprick tests were recorded bilaterally on the cheek before, during and after stimulation. If the stimulation induced a CH, the patients could treat with the SPG stimulator, and/or with their usual acute therapy (sumatriptan injection or oxygen). In addition, a peripheral venous catheter was inserted into the left or right antecubital vein for the drawing of blood samples. We collected samples at baseline (0 min), twice during stimulation (10 and 25 min) and after stimulation (45 min). One hour after the start of stimulation, the patient was discharged from the hospital. The patients were informed that SPG neurostimulation might induce headache or CH attacks in some individuals, but the timing and the characteristics of headache were not discussed.

All participants gave written informed consent to participate in the study, which was approved by the Ethics Committee of Copenhagen (H-15005609) and the Danish Data Protection Agency, and conducted accordance to the Helsinki II declaration of 1964, as revised in 2013. The study was registered at ClinicalTrials.gov (NCT02510729).

Cluster-like attack

To define criteria for an experimentally-induced cluster-like attack, it is important to notice: a) Patients are usually very familiar with their usual cluster attacks and are therefore able to report whether the induced attacks mimic their spontaneous attacks; b) most patients can predict the development of CH in the early stage of the attack, but for ethical reasons cannot be denied treatment. Therefore, we allowed patients to treat head pain before it became severe. The criteria for an experimentally-induced cluster-like attack are defined as the following:

Cluster-like headache attack fulfilling either (1) or (2):

Headache described as mimicking the patient’s usual cluster headache attack (with or without CAS)

Headache fulfilling criteria B and C for cluster headache according to the IHS ICHD-3 beta criteria:

Severe unilateral pain lasting 15–180 minutes

Either or both of the following:

at least one cephalic autonomic symptom (CAS), ipsilateral to the headache

a sense of restlessness

Headache intensity

Headache intensity was recorded at baseline and every 10 min up to 60 min on a numerical rating scale (NRS) from 0 to 10, where 0 is no headache, 1 represents a very mild headache (including an altered sensation in the head leading up to pain), 5 is headache of moderate intensity and 10 is the worst headache imaginable (8). Headache characteristics, CAS, whether headache mimicked usual CH, adverse events, rescue therapy (HF SPG stimulation, oxygen or sumatriptan) and its efficacy were also assessed. After discharge from the hospital, the patients were carefully instructed to record any cluster attacks until 24 h post-stimulation by a self-administered headache questionnaire similar to the one that was used in hospital.

Cephalic autonomic symptoms

We recorded the following CAS symptoms during the immediate phase (0–60 min) using a questionnaire and by observation: Nasal congestion, rhinorrhea, lacrimation, conjuctival injection, eyelid edema, ptosis, miosis, facial flushing and facial sweating. In addition, we recorded whether the patients experienced restlessness.

Mechanical perception and pain thresholds

We measured mechanical perception and pain thresholds to see whether sensory Aβ-fibers or C-fibers were affected during LF stimulation. We registered on both the ipsi- and contralateral side of stimulation on the cheek corresponding to the second branch of the trigeminal nerve (V2) at baseline, 15 min and 45 min after stimulation. An assessing investigator (ASP) was blind in respect to stimulation protocol.

The mechanical detection threshold was determined according to the standardized protocol from the German Research Network on Neuropathic Pain (DFNS) (9,10) by a set of 12 calibrated von Frey hairs (VFH; Marstock, Marburg, Germany; diameter 0.4–0.65 mm, force 0.25–512 mN). Beginning with a force of 16 mN, which each subject felt, the next lower hair was applied until the subject no longer felt the stimulus. Then, the next higher hair was applied until the subject noticed the stimulus. Using the method of limits, five infra- and five suprathreshold values were obtained and the geometrical mean was taken as the individual mechanical detection threshold. The mechanical pain threshold was tested with a series of seven standardized pinprick probes (diameter 0.2 mm, force 8–256 mN). The lowest force (8 mN) was applied first and was followed by the next higher force until the subject signaled a pricking pain. According to the method of limits, five infra- and five suprathreshold values were obtained and the geometric mean was considered as the individual mechanical pain threshold.

Blood sampling

Blood was drawn from a venous catheter into a 20 mL plastic syringe. The catheter was flushed with isotonic saline after each sampling. The blood from the syringe was then transferred into a lithium heparin tube for VIP and an EDTA tube for PACAP38 measurements. Both tubes were pre-cooled and contained aprotinin (Trasylol). Subsequently, the tubes were inverted several times and were centrifuged at 4℃ at 1851 g for 10 min. Plasma were then transferred to a polypropylene tube and stored at −25℃ for later analysis. Plasma concentrations of PACAP38 and VIP were measured by validated radioimmunoassay, as previously described (11,12).

Vital variables

MAP and HR were measured every 10 min in-hospital using an auto-inflatable cuff (ProPac Encore®; Welch Allyn Protocol).

Statistical analysis

All values are presented as median values (with interquartile ranges), except for vital variables presented as mean values ± 95% confidence interval (CI). Baseline was defined as before stimulation (0 min). Immediate phase was defined as 0–60 min after start of stimulation and delayed phase as > 1–24 h.

Calculation of sample size was based on the difference between two paired groups reporting cluster-like attacks after LF and sham stimulation in the immediate phase (0–1 h), at 5% significance with 80% power. We assumed that 60% of patients would report an attack after LF and 25% after sham and estimated that inclusion of at least 16 subjects would be sufficient in a crossover study (http://biomath.info/power/ based on JL Fleiss et al., Statistical Methods for Rates and Proportions). This assumption was mainly based on placebo-controlled provocation studies in migraine patients (13 –15) due to lack of placebo- or sham-controlled provocation studies in CH patients.

Primary end-points were the difference in incidence of cluster-like attacks, CAS, and the difference in area under the curve (AUC) for headache intensity scores during the immediate phase between LF and sham stimulation. The secondary end-points were differences between LF and sham in number of cluster attacks during the delayed phase and the entire registration period (0–24 h). In addition, we also tested the difference in AUC for mean arterial pressure (MAP) and heart rate (HR), mechanical detection and pain thresholds and plasma concentrations of PACAP38 and VIP.

Incidence of cluster-like attacks, headache, CAS, and adverse events were analyzed as binary categorical data using McNemar’s test. We calculated AUC according to the trapezium rule (16) to obtain a summary measure and to analyze the differences in response between LF stimulation and sham. Baseline was subtracted before calculating AUC to reduce variation between sessions within each subject. Analysis of AUC values were performed with a paired two-way t-test, except headache scores, mechanical perception, and pain thresholds, which were tested with the Wilcoxon signed rank test. Differences in mechanical perception and pain thresholds before (0 min), during (15 min) and after stimulation (45 min) for LF and sham were also analyzed using the Friedman test.

The p-values were not adjusted for multiple testing. All analyses were performed with SPSS Statistics version 22 for Windows (IBM, Chicago, IL, USA) and GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA) for the figures. A p-value < 0.05 was considered the level of significance.

Results

Twenty CH patients completed the study (four females/16 males). The mean age was 48 (range 25–77 years) and the mean CH attack frequency was 19 attacks per month (range 0–30 days per month). Fifteen patients had chronic CH. One patient was excluded due to personal issues after completing the first study day (LF) and was not included in the statistical analysis. The clinical characteristics of patients who completed the study are described in Table 1. The average time between implantation of the Pulsante system and the first study day for the patients was 775 days (range: 169–1705 days). The amplitude used during LF stimulation was between 0.7–2.1 mA depending on the maximum strength tolerated by each patient without them feeling discomfort or pain.

Table 1.

Clinical characteristics of the 20 completed study patients.

Patient- ID	Age/ gender	Cluster history (years)	Implant side	Cluster attacks days/month	Average cluster attacks/day	Days since implant¹	Effect of SPG HF therapy²	Prophylactic medication
1	52/M	CCH (13)	Left	0	0	1649	Yes	None
2	77/M	CCH (18)	Left	30	1	592	Yes	Verapamil, gabapentin
3	49/M	CCH (26)	Left	30	3	1698	No	None
4	64/M	CCH (26)	Left	30	2	647	Yes	Verapamil
5	61/M	CCH (7)	Right	30	4	579	Yes	Verapamil, gabapentin
6	67/M	ECH (31)	Right	30	6	1705	Yes	None
7	25/F	ECH (11)	Left	15	0.5	739	Yes	None
8	30/M	CCH (8)	Right	30	8	169	No	None
9	48/M	ECH (14)	Right	0	0	596	Yes	Verapamil
10	54/F	CCH (27)	Right	15	0.5	1120	No	Verapamil
11	32/M	CCH (24)	Right	30	3	988	No	None
12	42/M	CCH (31)	Left	30	3	912	Yes	Verapamil, gabapentin
13	51/M	CCH (15)	Right	30	2	358	No	None
14	40/M	CCH (27)	Left	15	0.5	256	Yes	Pregabalin
15	30/F	ECH (14)	Right	4	0.14	809	Yes	None
16	54/M	CCH (7)	Left	15	0.5	459	Yes	Candesartan, amlodipine
17	60/M	CCH (3)	Left	4	0.14	634	Yes	None
18	29/M	ECH (15)	Right	4	0.14	466	Yes	None
19	40/F	CCH (5)	Left	4	2	403	Yes	None
20	33/M	CCH (15)	Left	30	4	711	Yes	Verapamil, topiramate
Average/ Total	48/ 16 M/4F	5 ECH 15 CCH	9 Right 11 Left	19	2	775	15 Yes 5 No	9 Yes 11 None

F: female; M: male; CCH: chronic cluster headache; ECH: episodic cluster headache; SPG: sphenopalatine ganglion; HF: high frequency (therapeutic stimulation).

Number of days between implantation of the Pulsante system and the first study day of the present study.

Self-reported by patients.

Two of the patients had participated in the previous provocation study with LF stimulation (7). Two of the patients were in remission and had not had cluster attacks in more than one month. Four patients suffered from chronic tension-type headache, while 16 patients were headache free at baseline on both experimental days. All patients experience autonomic symptoms during their usual cluster attacks. Eleven out of 20 patients (55%) reported premonitory symptoms (yawning, unusual fatigue, thirst, hunger, mood swings, stiff neck, or poor concentration) that forewarned their usual cluster attacks when they were asked about these symptoms on their first visit.

One patient on LF day and one patient on sham day had missing values in their Von Frey measurements, because of damaged Von Frey hairs. Data at these time points were excluded from the statistical analysis and an extrapolation was done.

Cluster-like attacks and headache

In the immediate phase (0–60 min), seven out of 20 (35%) patients reported a cluster-like attack after LF stimulation compared to five (25%) after sham (p = 0.724). The median time to onset of cluster-like attacks was 20 min (range 10–40 min) after LF and 30 min (range 20–40 min) after sham. Only one (5%) patient took acute rescue therapy (oxygen) in the immediate phase after LF, which was effective, whereas none took rescue therapy after sham (p = 1.000). One of the patients who were in remission had an immediate cluster attack after 10 min of LF stimulation.

Headache characteristics for all patients in the immediate phase are presented in Table 2. We found no difference in the immediate phase (AUC_0-60 min) for headache intensity between LF and sham stimulation (p = 0.294) (Figure 1). The median peak headache intensity in the immediate phase after both LF and sham stimulation was 0.5 (range 0–8).

Table 2.

Headache characteristics of 20 cluster patients in the immediate phase (0–60 min) after 30 min of SPG stimulation using LF or sham.

Patient-ID		Headache characteristics/ peak intensity	Autonomic symptoms/ Restlessness¹	Time to peak headache	Time to onset of CAS	Mimics usual cluster attack (onset)²	Acute rescue therapy (time)/effect³
1	LF	Left/2	None	30 min	NA	Yes (10 min)	None
Sham	Left/2	None	50 min	NA	No	None
2	LF	None	inj/lac	NA	10 min	No	NA
Sham	None	inj/lac/ede	NA	20 min	No	NA
3	LF	Left/2	inj/lac/rhi/con/ ede/pto/mio/res	40 min	10 min	Yes (30 min)	None
Sham	Left/1	lac/con/ede/pto/res	10 min	10 min	Yes (20 min)	None
4	LF	None	lac/ede/mio	NA	20 min	No	NA
Sham	Left/1	inj/lac/ede/pto/mio	40 min	10 min	Yes (40 min)	None
5	LF	None	res	NA	NA	No	NA
Sham	None	None	NA	NA	No	NA
6	LF	Left/1	rhi/con/ede/ pto/mio/res	30 min	10 min	No	None
Sham	None	pto	NA	10 min	No	NA
7	LF	None	None	NA	NA	No	NA
Sham	Left/4	pto/res	20 min	60	Yes (30 min)	None
8	LF	None	con/ede/pto	NA	10 min	No	NA
Sham	None	None	NA	NA	No	NA
9	LF	None	inj/lac/rhi/con/mio	NA	10 min	No	NA
Sham	Right/1	None	40 min	NA	No	None
10	LF	None	con/pto/res	NA	10 min	No	NA
Sham	None	None	NA	NA	No	NA
11	LF	Right/2	inj/lac/rhi/con/res	30 min	10 min	No	None
Sham	None	None	NA	NA	No	NA
12	LF	None	lac/rhi	NA	10 min	No	NA
Sham	None	None	NA	NA	No	NA
13	LF	Left/2	inj/lac/con/pto	40 min	10 min	Yes (40 min)	None
Sham	None	None	NA	NA	No	NA
14	LF	• Bilat/3	inj/rhi	NA	10 min	No	None
Sham	• Bilat/3	inj/lac	NA	20 min	No	None
15	LF	None	None	NA	NA	No	NA
Sham	None	None	NA	NA	No	NA
16	LF	Left/8	lac/rhi/con/ ede/pto/res	20 min	10 min	Yes (20 min)	Oxygen (20 min)/Yes
Sham	Left/8	pto/res	50 min	10 min	No	None
17	LF	Left/6	ede/pto	50 min	20 min	Yes (20 min)	None
Sham	Left/7	rhi/con/pto	20 min	20 min	No	None
18	LF	None	None	NA	NA	No	NA
Sham	Bilat/1	None	20 min	NA	Yes (30 min)	None
19	LF	Left/6	inj/lac/ede/res	30 min	10 min	Yes (20 min)	None
Sham	Left/4	inj/lac/mio/res	30 min	20 min	Yes (20 min)	None
20	LF	Bilat/4	lac/rhi/con/res	50 min	30 min	Yes (30 min)	None
Sham	None	None	NA	NA	No	NA

Autonomic symptoms/restlessness: Inj: conjuctival injection; lac: lacrimation; rhi: rhinorrhea; con: nasal congestion; ede: eyelid edema; pto: ptosis; mio: miosis; res: restlessness.

Cluster-like attack: Defined according to criteria described in methods

Acute therapy: Therapeutic SPG stimulation, oxygen or sumatriptan. Therapy effect: According to patient.

NA: not applicable

• Daily tension-type headache

Figure 1.

Median (red/blue thick line) and individual (thin lines) headache intensity on a 0–10 NRS for 20 CH patients after LF and sham stimulation (0–30 min) during the immediate phase (0–60 min). There was no difference in the AUC_{0-60 min} between LF and sham stimulation.

CAS and restlessness

In the immediate phase, 16 (80%) patients experienced CAS after LF stimulation, while nine (45%) reported CAS after sham (p = 0.046) (Table 3). The assessing blinded investigator also recorded these CAS. The median time for onset of CAS was 10 min (range: 10–30 min) after the start of LF stimulation and 20 min (10–60 min) after the start of sham stimulation (p = 0.424). Nine patients developed CAS after LF stimulation but not headache, whereas five patients reported CAS and no headache on sham day (Table 2).

Table 3.

Cephalic autonomic symptoms (CAS) and sense of restlessness reported by patients (n = 20) after low frequency (LF) and sham stimulation in the immediate phase (0–60 min).

	n = 20
Reported symptoms	LF	Sham	p-value¹
Any CAS (0–60 min)	16	9	0.046¹
Nasal congestion	9	2	0.046¹
Rhinorrhea	8	1	0.046¹
Lacrimation	10	5	0.131
Conjuctival injection	7	4	0.371
Eyelid edema	7	3	0.221
Ptosis	7	6	1.000
Miosis	4	2	0.617
Facial flushing	3	1	0.617
Facial sweating	0	0	1.000
Restlessness (0–60 min)	8	4	0.221

Data are shown as number of patients, ¹p-value: McNemar’s test

Mechanical perception and pain thresholds

We found no difference in mechanical detection or pain thresholds between LF and sham stimulation on both ipsi- and contralateral side (Figure 2). We also found no changes from baseline for LF or sham stimulation on both ipsi- and contralateral side (p > 0.05, the Friedman test, data not shown).

Figure 2.

Median with interquartile range (IQR) of the mechanical detection and pain thresholds (mN) measured by Von Frey hairs and pinprick, respectively, before (0 min), during (15 min) and after (45 min) LF or sham stimulation in 20 patients. We found no difference in AUC between LF and sham stimulation on both ipsi- and contralateral side. There was no difference in change from baseline for LF or sham stimulation on both sides (p > 0.05).

Mean arterial blood pressure and heart rate

No difference in mean arterial blood pressure and heart rate and was found between LF stimulation and sham during the immediate phase (MAP AUC_0-60 min, p = 0.351 and HR AUC_0-60 min, p = 0.419) (Figure 3).

Figure 3.

Mean changes ± 95% CI in MAP and HR after LF or sham stimulation during the immediate phase (0–60 min) in CH patients (n = 20). There was no difference in the AUC_{0-60 min} for MAP or HR between the two groups.

Delayed phase (>1 h–24 h)

In the delayed phase, 14 (70%) patients reported an attack after LF stimulation compared to 15 (75%) after sham (p = 1.000). Ten (50%) patients took rescue medication after LF stimulation compared to 14 (70%) after sham (p = 0.751). Fourteen (70%) patients reported CAS after LF stimulation, while 15 (75%) reported CAS after sham (p = 1.000).

During the entire registration period (0–24 h), 11 patients reported > 1 attack after LF, whereas 10 patients had > 1 attack after sham (Figure 4). The median time of treatment (Pulsante™, oxygen or sumatriptan) was 4 h (range 0.66–22 h) after LF and 7 h (range 1.5–21.5 h) after sham (p = 0.612). All patients who took rescue therapy reported it to have effect for at least one of their attacks.

Figure 4.

Median number of cluster-like attacks (with quartiles) reported by patients (n = 20) after LF and sham during the entire registration period of 24 h (immediate phase + delayed phase) after stimulation. p-value: Wilcoxon signed rank test.

Blood levels of PACAP38 and VIP

We found no difference in AUC for plasma concentrations of VIP or PACAP38 between LF and sham stimulation (p > 0.05) (Figure 5).

Figure 5.

Plasma concentrations of VIP and PACAP38 (median with interquartile ranges) before (0 min), during (10 and 25 min) and after (45 min) LF or sham stimulation in 20 patients. We found no difference in AUC between LF and sham stimulation for VIP (p = 0.373) or PACAP38 (p = 0.723). P_AUC-value: Paired t-test.

Sub-analysis of patients with effect of SPG therapy

Of the 20 patients, 15 had effects from high frequency SPG therapy (Table 1). We did an explorative analysis of these 15 patients to see whether they had a more apparent response in headache intensity, QST or incidence of CAS compared to patients with no effect of SPG therapy and sham stimulation. We found no associations between effect of SPG therapy and any of the variables (p > 0.05).

Discussion

The major outcome of the present study is that LF stimulation of SPG induced more CAS compared to sham (80% vs. 45%). However, LF stimulation failed to induce more cluster-like attacks compared to sham (35% vs. 25%) in the immediate phase (0–60 min). The use of rescue therapy was low, suggesting that LF-provoked attacks were less severe than is typical for spontaneous cluster attacks. Explorative analyses revealed no associations between response in headache after LF stimulation and effect of HF SPG therapy. These data suggest that increased cranial parasympathetic outflow per se is probably not sufficient to induce a cluster-like attack.

In our previous study, LF neurostimulation of the SPG induced immediate cluster-like headache attacks in three out of six patients with chronic CH (7). The differences between the two studies include: a) 5 Hz as LF stimulation with a duration of 3 min (7) in contrast to 20 Hz over 30 min, b) HF stimulation (120 Hz) as “sham” stimulation (7), which could theoretically have prevented a nocebo effect on sham HF experimental days. Notably, we found a relatively high incidence of attacks (five out of 20 patients) after sham stimulation, suggesting the nocebo effect (17). Physiologically, LF stimulation between 7 and 45 min of the SPG (10–20 Hz) in animal studies has revealed dilation of intra- and extracranial arteries, increased cerebral blood flow and plasma protein extravasation in the dura mater (18 –22). We chose 20 Hz in the present study because we conducted three single-blinded pilot experiments in migraine patients showing that this frequency can cause rhinorrhea, decrease in mean blood flow velocity of the middle cerebral artery and induction of headache (unpublished data). We chose 30 min for the duration of stimulation, because we wanted to ensure a robust activation of the SPG. However, an animal study has shown that parasympathetic activation with 10 Hz peaked after 45 seconds of stimulation, at which point it decreased even with ongoing stimulation (23). It is therefore possible that using trains of pulses, e.g. 1 s on, 1 s off, could give a better effect (19).

In the present study, the majority of the patients (median: 641 days, range: 169–1705 days) underwent the SPG implant surgery more than two years ago. Sixteen patients still used the SPG stimulation, which might have resulted in the long-term effects of stimulation on neuroplasticity in the brain. Two patients reported remission for at least a month before the study. Interestingly, one of these patients developed a cluster-like attack after 10 min of LF stimulation that mimicked his usual cluster attacks. The other patient in remission did not develop an immediate cluster-like attack, but experienced CAS after LF stimulation. Since all patients participated in an ongoing registry study, we were not able to differentiate between patients who had efficacy of the SPG therapy or not, due to conflict of interest. However, explorative analysis of patients with self-reported effect of SPG HF therapy showed no stronger effect of LF stimulation compared to sham. Although SPG is the largest extracranial parasympathetic ganglion, sensory fibers from the maxillary nerve also project through the ganglion (24). It is therefore possible that SPG stimulation activates sensory Aβ-fibers or C-fibers. However, our data of mechanical perception and pain thresholds suggested no effect of SPG stimulation on Aβ-fiber (von Frey hairs test) or C-fiber (pinprick test) function. PACAP38 and structurally related VIP (25) are released from parasympathetic nerve fibers. Possible elevated plasma levels of both peptides may be used as a marker for parasympathetic activation during SPG stimulation. Neurostimulation of the vagal nerve has shown efferent release of VIP (26,27). Our data suggested that LF stimulation of the SPG did not affect plasma levels of PACAP38 or VIP. An exploratory study in five episodic CH patients reported elevated plasma PACAP38 collected from the cubital vein during CH attacks compared to the inter-bout phase (28). We believe that some important factors must be considered before using PACAP38 or VIP as parasympathetic biomarkers in patients: a) to what extent local release of parasympathetic peptides upon SPG stimulation can be detected in the peripheral circulation; b) when release of parasympathetic peptides reaches its peak concentration in the blood; c) assay validation should be a requirement.

Notably, most of the patients (75%) in the present study suffered from chronic CH. Frequent spontaneous attacks might therefore have influenced the results. In support, we found a high chance of spontaneous attacks with a mean of 19 attacks/month (range 0–30) and 2 attacks/day (range 0.14–10), and thereby a 63% chance of having an attack on any given day. The strength of the present study is the relatively large sample size and that we used a within-subjects design, where patients received two stimulation conditions, active and sham, over two sessions. In addition, it is notable that CAS were also observed by the same blinded investigator and not by the report of the patients alone.

Most patients are familiar with the Pulsante system, and the question is whether prior experience of SPG stimulation had a significant effect on patient-reported CH-like attacks. In addition, the patients would feel a difference between the two stimulations, as sham does not cause any sensation. In our protocol, all patients were told that stimulation frequencies would be different from those they used to treat spontaneous CH attacks. The term “sham stimulation” was neither included in informed consent or discussed with patients. The patients were informed that neurostimulation might induce headache or CH attacks in some individuals, but the timing and the characteristics of headache were not discussed. If the patients had become unblinded by the “active” stimulation, then it would be expected that such unblinding may have led to high incidence of CH-like attacks, since they would be more likely to believe that they were receiving attack provoking stimulation. The patients were randomly assigned into two groups: One group received sham then LF stimulation; the other group received LF stimulation first, then sham. The incidence of CH-like attack on the first experimental day was 5 out of 10 after LF stimulation and 3 out of 10 after sham. We conclude that there is no evidence that any unblinding that may have occurred in the present study was likely to have affected the outcome.

The present study indicates that the SPG activation is sufficient to induce CAS but cannot induce CH attacks, suggesting that both activation of sensory afferents and deep brain structures are necessary for initiation of a CH attack. Experimentally, direct stimulation of the superior salivatory nucleus causes cranial autonomic outflow and neuronal firing in the trigeminocervical complex (29), but activation of this pathway alone may not be sufficient to induce CH attacks. Thus, a positron emission tomography study showed activation of the ipsilateral inferior hypothalamic grey matter during the bout in CH patients (30) and there is a direct pathway from the hypothalamus to the superior salivatory nucleus and the SPG (31). In addition, CH attacks can only be induced by a lipophilic drug such as glyceryl trinitrate during a cluster bout, but not in the remission phase (32,33). Interestingly, there is also one case report in which a patient with complete trigeminal nerve section reported attacks with cranial autonomic symptoms, suggesting that this pathway can be activated from the brain without the trigeminal-autonomic reflex (34). Taken together, these findings suggest that parasympathetic efferent activation alone is insufficient to induce a cluster attack, and that activation of trigeminal afferents and deep brain structures including the hypothalamus are important for an initiation of cluster headache.

Conclusion

The present data suggest that parasympathetic outflow is sufficient to induce CAS but not sufficient to induce CH attacks in CH patients. We suggest that activation of deep brain structures and subsequent afferent pain input is important for the generation of attacks. The present data also raise the question of the mode of action of SPG stimulation in CH, and support the notion that neurostimulation modulates the higher order processing centers of the brain (35).

Article highlights

LF stimulation of the SPG induced autonomic symptoms, but no CH attacks.

These data suggest that increased parasympathetic outflow is not sufficient to induce CH attacks in patients.

Footnotes

Acknowledgements

The authors thank all the patients who participated in this study and the skillful technical assistance of Anita Hansen is gratefully acknowledged.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr. Song Guo has received a travel grant from the ATI. Dr. Anja Sofie Petersen, Henrik Winther Schytz and Jan Fahrenkrug have nothing to declare. Dr. Mads Barloese has been a consultant for ATI and received a travel grant. Anthony Caparso is Vice President, Research and Chief Scientist at Autonomic Technologies, Inc. Prof. Rigmor Jensen is an advisory board member for ATI, Allergan, Neurocore and Medotech, and has received honoraria and lectured for MSD, Pfizer, Allergan, Berlin-Chemie, and Norpharma. Professor Messoud Ashina is a consultant and/or scientific adviser/speaker for the Allergan, Alder, Amgen, ATI and Eli Lilly.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Further thanks to funding from Novo Nordisk Foundation (NNF11OC1014333), Independent Research-Medical Sciences (FSS) (DFF-1331-00210A), Lundbeck Foundation (R155-2014-171), FP7-EUROHEADPAIN (602633) and unrestricted grant from Autonomic Technologies, Inc.

References

Uddman

Tajti

Möller

et al.

Neuronal messengers and peptide receptors in the human sphenopalatine and otic ganglia. Brain Res 1999; 826: 193–199.

Goadsby

. Pathophysiology of cluster headache: A trigeminal autonomic cephalgia. Lancet Neurol 2002; 1: 251–257.

Goadsby

Lipton

. A review of paroxysmal hemicranias, SUNCT syndrome and other short-lasting headaches with autonomic feature, including new cases. Brain 1997; 120: 193–209.

Jürgens

Leone

. Pearls and pitfalls: Neurostimulation in headache. Cephalalgia 2013; 33: 512–525.

Magis

Schoenen

. Advances and challenges in neurostimulation for headaches. Lancet Neurol 2012; 11: 708–719.

Schoenen

Jensen

Lantéri-Minet

et al.

Stimulation of the sphenopalatine ganglion (SPG) for cluster headache treatment. Pathway CH-1: A randomized, sham-controlled study. Cephalalgia 2013; 33: 816–830.

Schytz

Barløse

Guo

et al.

Experimental activation of the sphenopalatine ganglion provokes cluster-like attacks in humans. Cephalalgia 2013; 33: 831–841.

Iversen

Olesen

Tfelt-Hansen

. Intravenous nitroglycerin as an experimental model of vascular headache. Basic characteristics. Pain 1989; 38: 17–24.

Rolke

Magerl

Campbell

et al.

Quantitative sensory testing: A comprehensive protocol for clinical trials. Eur J Pain 2006; 10: 77–88.

10.

Rolke

Baron

Maier

et al.

Quantitative sensory testing in the German Research Network on Neuropathic Pain (DFNS): Standardized protocol and reference values. Pain 2006; 123: 231–243.

11.

Birk

Sitarz

Petersen

et al.

The effect of intravenous PACAP38 on cerebral hemodynamics in healthy volunteers. Regul Pept 2007; 140: 185–191.

12.

Rahmann

Wienecke

Hansen

et al.

Vasoactive intestinal peptide causes marked cephalic vasodilation, but does not induce migraine. Cephalalgia 2008; 28: 226–236.

13.

Schytz

Birk

Wienecke

et al.

PACAP38 induces migraine-like attacks in patients with migraine without aura. Brain 2009; 132: 16–25.

14.

Antonova

Wienecke

Olesen

et al.

Prostaglandin E(2) induces immediate migraine-like attack in migraine patients without aura. Cephalalgia 2012; 32: 822–833.

15.

Guo

Olesen

Ashina

. Phosphodiesterase 3 inhibitor cilostazol induces migraine-like attacks via cyclic AMP increase. Brain 2014; 137: 2951–2959.

16.

Matthews

Altman

Campbell

et al.

Analysis of serial measurements in medical research. BMJ 1990; 300: 230–235.

17.

Mitsikostas

Mantonakis

Chalarakis

. Nocebo is the enemy, not placebo. A meta-analysis of reported side effects after placebo treatment in headaches. Cephalalgia 2011; 31: 550–561.

18.

Goadsby

. Sphenopalatine ganglion stimulation increases regional cerebral blood flow independent of glucose utilization in the cat. Brain Res 1990; 506: 145–148.

19.

Delépine

Aubineau

. Plasma protein extravasation induced in the rat dura mater by stimulation of the parasympathetic sphenopalatine ganglion. Exp Neurol 1997; 147: 389–400.

20.

Yarnitsky

Lorian

Shalev

et al.

Reversal of cerebral vasospasm by sphenopalatine ganglion stimulation in a dog model of subarachnoid hemorrhage. Surg Neurol 2005; 64: 5–11. discussion 11.

21.

Bar-Shir

Shemesh

Nossin-Manor

et al.

Late stimulation of the sphenopalatine-ganglion in ischemic rats: Improvement in N-acetyl-aspartate levels and diffusion weighted imaging characteristics as seen by MR. J Magn Reson Imaging 2010; 31: 1355–1363.

22.

Takahashi

Zhang

Z-D

Macdonald

. Sphenopalatine ganglion stimulation for vasospasm after experimental subarachnoid hemorrhage. J Neurosurg 2011; 114: 1104–1109.

23.

Suzuki

Hardebo

Kåhrström

et al.

Effect on cortical blood flow of electrical stimulation of trigeminal cerebrovascular nerve fibres in the rat. Acta Physiol Scand 1990; 138: 307–316.

24.

Moore KL, Agur AMR and Dalley AF. Clinically oriented anatomy. 7th ed. Baltimore: Lippincott Williams & Wilkins, 2013.

25.

Vaudry

Falluel-Morel

Bourgault

et al.

Pituitary adenylate cyclase-activating polypeptide and its receptors: 20 years after the discovery. Pharmacol Rev 2009; 61: 283–357.

26.

Björck

Fahrenkrug

Jivegård

et al.

Release of immunoreactive vasoactive intestinal peptide (VIP) from the gallbladder in response to vagal stimulation. Acta Physiol Scand 1986; 128: 639–642.

27.

Jensen

Aggestrup

Fahrenkrug

et al.

Electric stimulation of the autonomic nervous system and vasoactive intestinal peptide and substance P release on the lower esophageal sphincter. Acta Gastroenterol Latinoam 1990; 20: 131–136.

28.

Tuka

Szabó

Tóth

et al.

Release of PACAP-38 in episodic cluster headache patients – an exploratory study. J Head Pain 2016; 17: 69–69.

29.

Akerman

Holland

Summ

et al.

A translational in vivo model of trigeminal autonomic cephalalgias: Therapeutic characterization. Brain 2012; 135: 3664–3675.

30.

May

Bahra

Büchel

et al.

Hypothalamic activation in cluster headache attacks. Lancet 1998; 352: 275–278.

31.

Spencer

Sawyer

Wada

et al.

CNS projections to the pterygopalatine parasympathetic preganglionic neurons in the rat: A retrograde transneuronal viral cell body labeling study. Brain Res 1990; 534: 149–169.

32.

Ekbom

. Nitrolglycerin as a provocative agent in cluster headache. Arch Neurol 1968; 19: 487–493.

33.

Fanciullacci

Alessandri

Sicuteri

et al.

Responsiveness of the trigeminovascular system to nitroglycerine in cluster headache patients. Brain 1997; 120: 283–288.

34.

Matharu

Goadsby

. Persistence of attacks of cluster headache after trigeminal nerve root section. Brain 2002; 125: 976–984.

35.

Ambrosini

D’Alessio

Magis

et al.

Targeting pericranial nerve branches to treat migraine: Current approaches and perspectives. Cephalalgia 2015; 35: 1308–1322.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB

Cranial parasympathetic activation induces autonomic symptoms but no cluster headache attacks

Abstract

Background

Methods

Results

Conclusion

Study protocol

Keywords

Introduction

Materials and methods

Design

Cluster-like attack

Headache intensity

Cephalic autonomic symptoms

Mechanical perception and pain thresholds

Blood sampling

Vital variables

Statistical analysis

Results

Cluster-like attacks and headache

CAS and restlessness

Mechanical perception and pain thresholds

Mean arterial blood pressure and heart rate

Delayed phase (>1 h–24 h)

Blood levels of PACAP38 and VIP

Sub-analysis of patients with effect of SPG therapy

Discussion

Conclusion

Article highlights

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

References

Supplementary Material