Abstract
Background
Recently it has been suggested that low frequency stimulation of the sphenopalatine ganglion (SPG) may provoke cluster-like attacks in cluster headache (CH) patients. The question arises whether a robust activation of cranial autonomic symptoms is sufficient to trigger CH attacks.
Methods
Kinetic oscillation stimulation (KOS) of the nasal mucosa generates ipsilateral marked autonomic symptoms, among which lacrimation is quantitatively measurable. KOS was applied to 29 CH-patients, including both episodic and chronic course. We measured lacrimation at rest and during stimulation, and assessed CH attacks within 24 hours after the experiment.
Results
Autonomic symptoms including lacrimation were robust and significantly generated, compared to rest. Six patients were lost to follow-up, but did not develop an attack during their stay in the clinic. Of the remaining 23 patients, none developed an attack in the next 4 hours after stimulation, despite marked cranial autonomic symptoms during stimulation.
Discussion
Peripheral stimulation close to the SPG generated a strong parasympathetic response. However, this stimulation was not sufficient to induce CH attacks, which suggests that a central component is crucial to attack generation.
Keywords
Introduction
Autonomic symptoms ipsilateral to the affected side are one of the key features in CH, and include lacrimation, nasal congestion, rhinorrhea, facial/forehead sweating, myosis, ptosis, eyelid edema and conjunctival injection (1). It is thought that these symptoms are generated by the parasympathetic activation mediated by the superior salivatory nucleus (SSN) via the SPG. Clinically, the hypothalamus plays a crucial role in the initiation of CH attacks, which explains their circadian and circannual rhythmicity (2). Trigeminal neurons receive input from the SSN, which in turn receives input from the hypothalamus. Furthermore, it was proposed by Ivanusic and coworkers that parasympathetic neurons innervating the nasal mucosa and the lacrimal gland are directly modulated by 5-HT1D receptors in the SPG (3). The above-mentioned structures are thought to be the key components involved in the trigemino-autonomic reflex that becomes activated and is maintained, resembling a vicious circle during severe CH attacks (4).
Recently, it has been shown that high frequency stimulation of the SPG, which is the peripheral relay station for facial parasympathetic outflow, aborts CH attacks (5). The theoretical construct was that high frequency stimulation leads to a depolarization within the ganglion, which effectively blocks the peripheral parasympathetic innervation of the eye and nose. With this prior knowledge, Schytz and colleagues tested in an elegant paradigm whether it is possible to trigger CH attacks in patients by peripheral stimulation of the SPG at low frequency, which was suggested to activate the ganglion. They reported that 67% of the investigated CH patients developed a “cluster-like” attack following low frequency stimulation (6).
However, these results can be questioned, as some patients reported attacks as late as several hours after the stimulation, which calls into question whether they were triggered or indeed emerged spontaneously.
Thus, from a pathophysiological point of view, the following question is important: Is it possible to reliably trigger CH attacks by activating the peripheral parasympathetic nervous system? If so, the periphery plays a crucial role in attack generation, leaving the central aspect, namely the hypothalamic involvement, to be a mere change in state. The opposite point of view places the hypothalamus not only as a change in state but indeed as the coordinated initiator of both the pain and the autonomic symptoms (2).
Focusing on this question, we tested whether a robust initiation of cranial autonomic symptoms following peripheral stimulation of the nasal mucosa using the KOS paradigm is indeed sufficient to trigger CH attacks in patients in the active or even inactive period.
Material and methods
To investigate the effect of the KOS paradigm on triggering of CH attacks, 29 CH patients were recruited from the headache outpatient clinic of the University Medical Center Hamburg-Eppendorf (UKE). Written informed consent was obtained from all subjects, and the study was conducted according to the Declaration of Helsinki and approved by the Ethics Committee in Hamburg, Germany.
Experimental procedures
Characteristics of cluster headache patients and assessment of the next attack after KOS procedure at 68 Hz/95mbar.
M: male; f: female; eCHa: episodic cluster headache within an active period; cCH: chronic cluster headache; eCHi: episodic cluster headache within an inactive period; h: hour; d: day; w: week; mo: month; y: year; min: minute.
Kinetic-oscillation-stimulation (KOS) procedure
The KOS-device was used to mechanically stimulate the nasal mucosa and to elicit cranial autonomic symptoms that were quantifiable via lacrimation.
For the stimulation, the inflatable tip of the balloon catheter was wetted with medical paraffin and placed inside the nostril of the affected CH side. In order to stabilize the balloon catheter safely inside the nostril, the catheter was fixed with a clamp on a frame that was attached to the patient’s head (see (7)). Since mere manipulation, that is, placement of the catheter, caused an autonomic reaction including lacrimation, which lasted for 30–40 seconds, the stimulation was started 60 seconds later to enable the autonomic system to decline activation. For the stimulation over 5 minutes, the balloon was inflated to a pressure of 95 mbar and oscillated with a frequency of 68 Hz (see (7)). Lacrimation was documented every minute for 5 minutes during stimulation.
The patients were monitored so that the measurements could be cancelled immediately in case of a triggered CH attack. Oxygen was provided as emergency medication.
For quantification of lacrimation, a standardized Schirmer’s-II test was performed during rest and during stimulation. Patients were tested in a standardized illuminated laboratory room. Sterile tear strips (Avimed GmbH, Wiesbaden, Germany) were placed inside the lower eyelid and patients were instructed to leave their eyes open and to fixate on a point in front of them during measurements for rest and stimulation conditions, in order to prevent the filter strip from slipping out of the eye.
During rest, lacrimation was assessed for 5 minutes and documented at 2 minutes, 4 minutes and 5 minutes for both eyes after the start of the experiment. The KOS procedure was performed subsequently to the resting condition.
Statistical analysis
The mean lacrimation difference (in mm) after 2 minutes between stimulation and rest was calculated. Statistical analysis of the data acquired in the Schirmer’s-II test was calculated in SPSS (version 21, IBM, USA). First, a Kolmogorov-Smirnov test was performed to test for normal distribution of the data. Subsequently the data was analyzed in a one sample t-test.
Follow-up
A follow-up telephone call was implemented within the following days after the experiment. Patients were asked when they experienced the next attack subsequently to the stimulation. Additionally, any discomfort concerning the eye or nose were inquired about.
Results
The KOS procedure evoked significantly (p < .0001) stronger lacrimation during stimulation compared to the resting condition. The lacrimation was accompanied by other cranial autonomic symptoms in most of the patients, including conjunctival injection and rhinorrhea on the stimulated side; however, those symptoms were not quantified. The mean lacrimation acquired during rest was 13.8 (±1.86) mm. The mean lacrimation during stimulation of the affected CH side was 24.1 (±1.97) mm (Figure 1).
Mean lacrimation (mm) during rest compared with mean lacrimation during stimulation measured on the affected CH-side, assessed 2 minutes after start of the measurement. Stimulation was performed at 68Hz/95mbar. Error bars show the standard error of the mean (SEM). A significantly stronger lacrimation was observed during stimulation (p<.0001), indicated by the asterisk.
Six of the patients were lost to follow-up. None of them developed an attack during the experimental procedure up to 15 minutes thereafter (the time patients stayed in the clinic). One patient reported an attack 4 hours after stimulation. None of the remaining patients developed a cluster headache or indeed any headache attack in the next 7 hours (for demographic details see Table 1).
Discussion
Kinetic oscillation stimulation (KOS) of the nasal mucosa generated a significant autonomic output (lacrimation) on the stimulated side, compared to rest. We were not able to trigger CH attacks in 23 CH patients.
The earliest CH attack reported in the present study started approximately 4 hours after stimulation procedure. This patient (no. 8) suffered from a relatively high attack frequency (3/day) and also experienced an attack the day before the study (see Table 1). Patient 7 experienced a CH attack approximately 7 hours after stimulation and also stated a relatively high attack frequency within active periods (4/day). Since both patients were in an active CH period, it is likely that these patients spontaneously developed an attack independent from KOS stimulation due to their high attack frequency.
One patient in the episodic inactive CH group (no. 29) developed a CH attack approximately 7 hours post stimulation. This patient stated that he had not experienced any attack within 5 months before the study, being out of the bout. Since experiments including the person concerned were performed during spring, it might be possible that this patient was at the brink of an active bout. Previous studies suggest that triggering CH attacks with nitroglycerine in eCH patients is only possible within their active period (8). The study by Schytz and coworkers did not include inactive CH patients, therefore it remains unclear if it is at all possible to initiate CH attacks by peripheral stimulation of the system via SPG low frequency stimulation in patients outside their active period (6). Since we tested a new paradigm, we decided to include also inactive episodic CH patients. Our data, however, strongly suggest that it is not possible to trigger CH attacks in patients, independent of whether they are in the active or inactive CH state.
Unfortunately, for six out of 29 CH patients, no follow-up data was available. It cannot be excluded that these patients developed CH attacks after they left the clinic. Nonetheless, we were not able to trigger a CH attack in any of the patients during their stay in the clinic, and we did not observe attacks within at least 4 hours following stimulation in the majority of the patients. It can be assumed that the likelihood of developing an attack is dependent on the overall attack frequency per day of the individual patient, and might be independent from the general classification of chronic and episodic course. We note that the headache frequency was rather low in most patients that were investigated. We assume that patients with a generally high attack frequency per day were rather concerned about triggering attacks and refused to take part in this study. It is theoretically possible that the elicited cranial parasympathetic output provoked by the KOS paradigm differs from the parasympathetic output that is generated during an attack, and that our stimulus was not sufficient. Given that we elicited a significant and robust parasympathetic output, which was comparable to the patient’s usual attacks, we do not think that this is the case. Our data strongly suggest that a mere peripheral cranial parasympathetic activation is not sufficient to evoke cluster attacks, as has been recently suggested (6). This raises the question of whether well-known triggers in cluster headache such as alcohol, histamine or NO-donors evoke attacks by yet-unknown central effects rather than mere vasodilation.
Conclusion
Taken together, we were not able to trigger CH-attacks following peripheral provocation of cranial autonomic symptoms in a cohort of 23 CH patients. These findings strongly suggest that a peripheral stimulation of trigeminal afferents and consequent parasympathetic output via the trigemino-parasympathetic reflex (9) in the nasal mucosa is not sufficient to trigger CH attacks. Therefore, we propose that the involvement of a central structure, namely the hypothalamus, is necessary for the initiation process of CH attacks.
Clinical implications
Kinetic oscillation stimulation (KOS) of the nasal mucosa results in robust autonomic activation ipsilateral to the stimulated side. Peripheral autonomic activation does not trigger CH attacks in patients. Investigating the trigemino-autonomic reflex could support a detailed understanding of the autonomic symptoms commonly observed in CH patients.
Footnotes
Acknowledgements
The authors thank the patients for helping with this study.
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: AM has been consultant and speaker for ATI in the past.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by an unrestricted scientific grant by Chordate Medical.