MRI of Intracranial Arteries in Nitroglycerin Induced Cluster Headache Attacks
Jan Hannerz, M.D.1 and Dan Greitz, M.D.2
Department of Neurology1 and Neuroradiology2 Karolinska hospital, Stockholm, Sweden Address all correspondence to: Jan Hannerz, M.D., Department of Neurology, Karolinska Hospital, S-104 01 Stockholm, Sweden Accepted for publication: July 26, 1992 SYNOPSIS
Eight patients with episodic cluster headache, five in active episode, three out of episode, were investigated as to diameters of intracranial arteries before and after nitroglycerin (NG) administration. The diameter of all intracranial carotids were increased about 10 minutes after NG, although more in the patients in episode than in patients out of episode. The dilatation remained for the next 60 minutes in the patients who did not get a cluster headache attack. There was a normalization of the diameters of the internal carotid arteries compared to the initial values, at maximum pain in all patients who got a cluster headache attack. Similar changes were also found in the basilar arteries. The findings support the hypothesis of a constriction of intracranial arteries at maximum pain in cluster headache attacks to stop the pain. Key words: cluster headache, MRI. (Headache 1992; 32:485-488) INTRODUCTION
It has been suggested that episodic cluster headache attacks are due to venous vasculitis in the middle fossa causing, in addition to deficient venous drainage, a local sympathicoplegia.1 Cluster headache attacks are known to appear in relation to REM sleep and after alcohol ingestion, both situations accompanied by changes in cerebral blood flow.2,3 Due to regional loss of regulation of arterial blood flow to the cavernous sinus, the impaired venous drainage of the region should be accentuated with increases of cerebral blood flow, which may in turn result in intense ipsilateral pain, conjunctival injection, lacrimation and rhinorrhea. The intensity of the pain has been suggested as a mechanism to induce constriction of arteries caudal to the sympathicoplegia, decreasing the blood flow to the phlebopathic area, normalizing the drainage of the middle fossa and ending the attack.1 Findings in a recent study using duplex ultrasound blood flow techniques in the common carotid arteries in cluster headache patients and controls support such a suggestion and indicate that in cluster headache attacks the intensity of the pain was related to decreased blood flow in both common carotid arteries.4 Nitroglycerin (NG) is known to provoke cluster headache attacks when a patient is in an active cluster period.3,5 The aim of this study was to investigate the diameters of extracranial and intracranial arteries caudal to the suggested sympathicoplegia in the cavernous region, before and after a NG provoked cluster headache attack and to test the hypothesis that constriction of arteries terminates an attack of cluster headache. The studies were performed with MRI in cluster headache patients who were in cluster period and in such patients who were out of cluster period as controls. PATIENTS AND METHODS
Eight patients with cluster headache according to the criteria of IHS, five in and three out of a cluster period, were investigated. The patients were studied when they had been free from cluster headache attacks for at least eight hours and free from medicine for the last 24 hours. They all received 1 mg nitroglycerin sublingually as provocation. The patients were investigated with MRI (Siemens T 1,0). All patients had the head fixed to the table so as to be able to study identical areas of the vessels with each examination.6 The internal carotid and vertebral arteries were studied extracranially at the level of the second cervical vertebra while the internal carotid and the basilar arteries were studied intracranially but at the extradural part of the internal carotid artery, i.e., about 5 mm caudal to the horizontal part of the carotid siphon. The studies were performed before and ten minutes after nitroglycerin provocation in all patients, at maximal pain and after the cluster headache attack was over, or 60 minutes after nitroglycerin administration in patients who did not develop an attack. Sagittal and coronal diameters of the intracranial arteries were measured and the mean of these two measurements was considered as the diameter. Changes in the extracranial vessels were only estimated as to an increase or a decrease compared to the diameter before the provocation. The data was analyzed for statistical significance using analysis of variance (ANOVA). RESULTS
Effects of NG administration on intracranial vessels. The diameter of all intracranial ca-
rotid arteries increased, compared to the initial size, ten minutes after NG administration. The diameter of all basilar arteries in patients who were in an active cluster headache period also increased compared to the initial size, but only one of the two studied patients who were out of the active cluster headache phase had a similar increase. There was no change of the diameter in the other patient. In the three patients who did not develop a cluster headache attack, the dilatation remained for more than 60 minutes after NG administration (Table 1, Figure 1). Pain related changes in intracranial vessels. All patients who developed a cluster headache attack showed a decrease in the diameters of both the internal carotid arteries and the basilar arteries at maximum pain, compared to the diameters recorded ten minutes after NG administration. At maximum pain the diameters usually returned to about the sizes which were found before NG administration. The diameter of the arteries remained largely the same after the headache attack was over. (Table 1, Figure 2). Effects of NG administration on extracranial vessels. There was dilatation of extracranial vessels, compared to the initial diameters, ten minutes after NG administration although not so pronounced as in the intracranial vessels. There was no systematic change in the diameters of the extracranial vessels in relation to maximum pain of the cluster headache attacks. DISCUSSION
It has been suggested that episodic cluster headache is due to a temporary venous vasculitis in the cavernous sinus causing a unilateral sympathicoplegia.1,4,7 Nitroglycerin is known to provoke attacks of cluster headache in patients who are in an active cluster period.3,5 The aim of this study was to investigate the intracranial arteries caudal to the cavernous sinus region, before, during and after a NG provoked cluster headache attack. In this study NG was found to dilate intracranial arteries in both cluster headache patients in and out of a cluster period; a dilatation which lasted for more than 60 minutes if no cluster headache attack was provoked. The average increase in the diameter of internal carotid arteries, compared to initial values, 10 minutes after NG administration was 34% in cluster headache patients in an active period and 14.9% for cluster headache patients out of an active period. The corresponding increase 15 minutes after
A Case 1 Before NG 10 min after NG Max pain 10 End of Pain
5,0 6,0 5,2 5,2
B 4,5 6,2 6,1 6,5
Table 1 C 3,5 5,2 3,4 3,9
Case 2
Case 5 Before NG 10 min after NG Max pain 5 End of pain
A
B
C
4,2 5,9 4,2 4,2
5,1 6,7 5,0 5,0
2,4 3,1 2,4 2,4
5,0 5,3 5,5
4,2 4,8 5,0
3,4 3,4 3,2
5,3 6,2 6,1
5,0 6,1 5,5
3,6 4,2 3,7
5,0 n.d. 5,5
4,2 n.d. 5,9
2,3 n.d. 2,3
Case 6
Before NG 10 min after NG Max pain 10 End of pain
4,2 5,1 4,5 4,2
n.d. n.d. n.d. n.d.
2,5 3,9 3,5 3,5
Case 3
Before NG 10 min after NG 60 min after NG
Case 7
Before NG 10 min after NG Max pain 6 End of pain
4,9 5,8 4,3 5,0
4,9 6,8 4,1 4,9
2,9 4,1 3,3 3,3
Case 4
Before NG 10 min after NG 60 min after NG
Case 8
Before NG 10 min after NG Max pain 6 End of pain
4,1 4,7 4,3 4,4
3,6 5,9 4,6 4,6
2,9 4,2 3,5 3,6
Before NG 10 min after NG 60 min after NG
A=Right internal carotid artery, B=Left internal carotid artery, C=Basilar artery, Pain Scale 0-10.
NG administration of the common carotid arteries in a recent ultrasound duplex study4 was 4.5 % in cluster headache patients in a cluster period and 12.4 % in patients out of a cluster period and 11.7 % in controls. Although this study only dealt with a few patients, which makes definite conclusions uncertain, there was a statistically significant difference (p
Jan Hannerz, M.D.1 and Dan Greitz, M.D.2
Department of Neurology1 and Neuroradiology2 Karolinska hospital, Stockholm, Sweden Address all correspondence to: Jan Hannerz, M.D., Department of Neurology, Karolinska Hospital, S-104 01 Stockholm, Sweden Accepted for publication: July 26, 1992 SYNOPSIS
Eight patients with episodic cluster headache, five in active episode, three out of episode, were investigated as to diameters of intracranial arteries before and after nitroglycerin (NG) administration. The diameter of all intracranial carotids were increased about 10 minutes after NG, although more in the patients in episode than in patients out of episode. The dilatation remained for the next 60 minutes in the patients who did not get a cluster headache attack. There was a normalization of the diameters of the internal carotid arteries compared to the initial values, at maximum pain in all patients who got a cluster headache attack. Similar changes were also found in the basilar arteries. The findings support the hypothesis of a constriction of intracranial arteries at maximum pain in cluster headache attacks to stop the pain. Key words: cluster headache, MRI. (Headache 1992; 32:485-488) INTRODUCTION
It has been suggested that episodic cluster headache attacks are due to venous vasculitis in the middle fossa causing, in addition to deficient venous drainage, a local sympathicoplegia.1 Cluster headache attacks are known to appear in relation to REM sleep and after alcohol ingestion, both situations accompanied by changes in cerebral blood flow.2,3 Due to regional loss of regulation of arterial blood flow to the cavernous sinus, the impaired venous drainage of the region should be accentuated with increases of cerebral blood flow, which may in turn result in intense ipsilateral pain, conjunctival injection, lacrimation and rhinorrhea. The intensity of the pain has been suggested as a mechanism to induce constriction of arteries caudal to the sympathicoplegia, decreasing the blood flow to the phlebopathic area, normalizing the drainage of the middle fossa and ending the attack.1 Findings in a recent study using duplex ultrasound blood flow techniques in the common carotid arteries in cluster headache patients and controls support such a suggestion and indicate that in cluster headache attacks the intensity of the pain was related to decreased blood flow in both common carotid arteries.4 Nitroglycerin (NG) is known to provoke cluster headache attacks when a patient is in an active cluster period.3,5 The aim of this study was to investigate the diameters of extracranial and intracranial arteries caudal to the suggested sympathicoplegia in the cavernous region, before and after a NG provoked cluster headache attack and to test the hypothesis that constriction of arteries terminates an attack of cluster headache. The studies were performed with MRI in cluster headache patients who were in cluster period and in such patients who were out of cluster period as controls. PATIENTS AND METHODS
Eight patients with cluster headache according to the criteria of IHS, five in and three out of a cluster period, were investigated. The patients were studied when they had been free from cluster headache attacks for at least eight hours and free from medicine for the last 24 hours. They all received 1 mg nitroglycerin sublingually as provocation. The patients were investigated with MRI (Siemens T 1,0). All patients had the head fixed to the table so as to be able to study identical areas of the vessels with each examination.6 The internal carotid and vertebral arteries were studied extracranially at the level of the second cervical vertebra while the internal carotid and the basilar arteries were studied intracranially but at the extradural part of the internal carotid artery, i.e., about 5 mm caudal to the horizontal part of the carotid siphon. The studies were performed before and ten minutes after nitroglycerin provocation in all patients, at maximal pain and after the cluster headache attack was over, or 60 minutes after nitroglycerin administration in patients who did not develop an attack. Sagittal and coronal diameters of the intracranial arteries were measured and the mean of these two measurements was considered as the diameter. Changes in the extracranial vessels were only estimated as to an increase or a decrease compared to the diameter before the provocation. The data was analyzed for statistical significance using analysis of variance (ANOVA). RESULTS
Effects of NG administration on intracranial vessels. The diameter of all intracranial ca-
rotid arteries increased, compared to the initial size, ten minutes after NG administration. The diameter of all basilar arteries in patients who were in an active cluster headache period also increased compared to the initial size, but only one of the two studied patients who were out of the active cluster headache phase had a similar increase. There was no change of the diameter in the other patient. In the three patients who did not develop a cluster headache attack, the dilatation remained for more than 60 minutes after NG administration (Table 1, Figure 1). Pain related changes in intracranial vessels. All patients who developed a cluster headache attack showed a decrease in the diameters of both the internal carotid arteries and the basilar arteries at maximum pain, compared to the diameters recorded ten minutes after NG administration. At maximum pain the diameters usually returned to about the sizes which were found before NG administration. The diameter of the arteries remained largely the same after the headache attack was over. (Table 1, Figure 2). Effects of NG administration on extracranial vessels. There was dilatation of extracranial vessels, compared to the initial diameters, ten minutes after NG administration although not so pronounced as in the intracranial vessels. There was no systematic change in the diameters of the extracranial vessels in relation to maximum pain of the cluster headache attacks. DISCUSSION
It has been suggested that episodic cluster headache is due to a temporary venous vasculitis in the cavernous sinus causing a unilateral sympathicoplegia.1,4,7 Nitroglycerin is known to provoke attacks of cluster headache in patients who are in an active cluster period.3,5 The aim of this study was to investigate the intracranial arteries caudal to the cavernous sinus region, before, during and after a NG provoked cluster headache attack. In this study NG was found to dilate intracranial arteries in both cluster headache patients in and out of a cluster period; a dilatation which lasted for more than 60 minutes if no cluster headache attack was provoked. The average increase in the diameter of internal carotid arteries, compared to initial values, 10 minutes after NG administration was 34% in cluster headache patients in an active period and 14.9% for cluster headache patients out of an active period. The corresponding increase 15 minutes after
A Case 1 Before NG 10 min after NG Max pain 10 End of Pain
5,0 6,0 5,2 5,2
B 4,5 6,2 6,1 6,5
Table 1 C 3,5 5,2 3,4 3,9
Case 2
Case 5 Before NG 10 min after NG Max pain 5 End of pain
A
B
C
4,2 5,9 4,2 4,2
5,1 6,7 5,0 5,0
2,4 3,1 2,4 2,4
5,0 5,3 5,5
4,2 4,8 5,0
3,4 3,4 3,2
5,3 6,2 6,1
5,0 6,1 5,5
3,6 4,2 3,7
5,0 n.d. 5,5
4,2 n.d. 5,9
2,3 n.d. 2,3
Case 6
Before NG 10 min after NG Max pain 10 End of pain
4,2 5,1 4,5 4,2
n.d. n.d. n.d. n.d.
2,5 3,9 3,5 3,5
Case 3
Before NG 10 min after NG 60 min after NG
Case 7
Before NG 10 min after NG Max pain 6 End of pain
4,9 5,8 4,3 5,0
4,9 6,8 4,1 4,9
2,9 4,1 3,3 3,3
Case 4
Before NG 10 min after NG 60 min after NG
Case 8
Before NG 10 min after NG Max pain 6 End of pain
4,1 4,7 4,3 4,4
3,6 5,9 4,6 4,6
2,9 4,2 3,5 3,6
Before NG 10 min after NG 60 min after NG
A=Right internal carotid artery, B=Left internal carotid artery, C=Basilar artery, Pain Scale 0-10.
NG administration of the common carotid arteries in a recent ultrasound duplex study4 was 4.5 % in cluster headache patients in a cluster period and 12.4 % in patients out of a cluster period and 11.7 % in controls. Although this study only dealt with a few patients, which makes definite conclusions uncertain, there was a statistically significant difference (p
