Cluster headache (CH) is characterized by a series of sudden attacks of short-lasting severe headache pain with ipsilateral autonomic features, including lacrimation, rhinorrhea, localized sweating, eyelid edema, and partial or complete Horner's syndrome. Just like in migraine, brush allodynia has been described for CH1 and for short lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT) syndrome.2 Administration of normobaric oxygen is part of the standard therapy for CH attacks. Here, we describe a young male with a first CH attack and the influence of oxygen on pain and concomitant cutaneous allodynia (CA).
Cluster headache (CH) is characterized by a series of sudden attacks of short-lasting severe headache pain with ipsilateral autonomic features, including lacrimation, rhinorrhea, localized sweating, eyelid edema, and partial or complete Horner's syndrome. Just like in migraine, brush allodynia has been described for CH and for short lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT) syndrome. Administration of normobaric oxygen is part of the standard therapy for CH attacks. Here, we describe a young male with a first CH attack and the influence of oxygen on pain and concomitant cutaneous allodynia.
A 49-year-old man experienced for the first time a right-sided severe stabbing headache, located around the right eye, right temple, and ear, beginning between 3 and 4 am. He had a sense of agitation and walked around his bedroom with the windows wide open. The attack lasted 45 minutes (without treatment), ceased, and recurred 3 times that night. The next night the attack started again at approximately the same time and in the same area of the head. Because the area of the cheek ipsilateral to the affected pain side was extremely sensitive when touched and the right eye was tearing, he looked in the mirror and noticed a swelling of the right eyelid with conjunctival injection. When he admitted himself to our headache center, he had no pain and clinical examination also showed no neurological deficit. A diagnosis of CH was made. Brain magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) of the intracranial and extracranial carotid circulation were performed to rule out a secondary CH; these were negative. MRI of the neck area and the cervical spine also showed no pathology.
The following night at 3:15, he again felt the same headache (intensity 10 on a verbal rating scale of 0-10) and was tested for the presence of CA. The same examiner tested for brush allodynia by lightly stroking his scalp and entire face region including the 3 sensory divisions of the trigeminal nerve sequentially and bilaterally with a small brush. Furthermore, an ice cube was used to test for cold allodynia. Brush-related CA was noted at the right temple, in all 3 sensory divisions of the trigeminal nerve, extending to the neck to the area of C2 and C3. The most intense allodynia occurred at the right temple at an intensity of 6 on a verbal scale of unpleasantness from 0 to 10. When the patient underwent oxygen therapy (100%, 14 L/minute), the pain attack ceased within 20 minutes. Although oxygen therapy was discontinued, CA persisted until the next headache attack started 40 minutes later. Oxygen therapy was recommenced immediately. Within 10 minutes, CH and CA resolved completely and did not start up again that night.
This first cluster episode lasted 2 months. The patient had 2 to 3 attacks each night. Treatment with 100% oxygen (14 L/min) stopped pain within 20 minutes and CA within half an hour. Between the attacks, the patient had no evidence of CA.
This report describes the development of CA during a CH attack in a reliable patient. Spatially, it affected the pain area at the ipsilateral face, before it affected the ipsilateral neck. To our knowledge, this is the first report of a patient having typical episodic CH with CA, in which both symptoms, namely pain and CA, completely resolved under therapy with 100% oxygen.
Cutaneous allodynia is the perception of pain when a nonnoxious stimulus is applied to normal skin. It has been extensively studied in patients with migraine, providing a theoretical basis for the assumption that neurons in the trigeminovascular system are sensitized in both diseases, thus causing CA and pain.3,4 This form of sensitization could be initiated by a prolonged barrage of incoming signals from nociceptors5 that could possibly activate nociceptive neurons and glial cells in the spinal cord.6 Sensitization of peripheral nociceptors that innervate intracranial blood vessels and the meninges may explain how mild mechanical stimuli such as brushing or slight pressure can aggravate the pain.
In accordance with a previous case report, we also found that brush as well as cold allodynia and thus central sensitization were marked by abrupt onset and termination.2 However, the suggestion that central sensitization of central-order trigeminovascular neurons may start shortly after pain initiation without becoming clinically detectable as CA until about 1 hour later can possibly be observed in migraine but not in trigemino-autonomic headaches. In CH and SUNCT, it seems that CA is attack- and syndrome-dependent.
One of the first studies to deal with a potential effectiveness of oxygen therapy in treating CH was conducted by Kudrow.7 While the benefits of oxygen therapy are apparent and proven in about 70% of patients, how and why it works remains unclear.8 The effect of normobaric oxygen on pain can be explained by a vasoconstrictor effect, which reduces the pressure and inflammation associated with CH. The mechanism by which oxygen alters sensory perception, as a sign of secondary sensitization of pain-processing, second-order neurons due to nociceptive inflow during the attack, is completely unknown. One potential explanation for this oxygen effect could be the stimulation of serotonin synthesis, possibly involved in the initiation and maintenance of CA.
Our case report shows that under oxygen therapy the pain attack ceased within 20 minutes and CA persisted until the next attack. Therefore, we might conclude that the cessation of pain per se did not diminish CA, which stands in contrast to the observation made by Rozen et al.2 In the next attack, which started 40 minutes later, pain and CA ceased nearly simultaneously thanks to continuous oxygen supply. A study by Schuh-Hofer et al showed hyperoxia to have an effect on neurogenic plasma protein extravasations, but not normoxia.9 Similar results were shown in Goadsby's work based on treatment with hyperoxia and the interaction of hyperoxia and neuropeptide release.10 Since we know that calcitonin-gene-related-peptide levels decreased during treatment with oxygen, we can suppose that the interaction of hyperoxia and neuropeptide release from trigeminal afferents also alters CA. Further studies must focus on acute treatment of patients with episodic CH and CA using normobaric oxygen. Furthermore, investigations should draw their attention on the discussion of a possible temporal profile of CA during a CH attack. For this reason, it would be interesting to test CH patients for CA at different time points during an attack, because this was not performed in our case study. Thus, all the more, one study shows no CA in CH patients,11 meaning that this issue is still debatable.
Acknowledgments: We would like to thank Prof. Burstein for his generous improvement and prolific discussion.