Persistence of attacks of cluster headache after trigeminal nerve root section

Cluster headache is a strictly unilateral headache that occurs in association with cranial autonomic features. We report a patient with a trigeminal nerve section who continued to have attacks. A 59-year-old man described a 14-year history of left-sided episodes of excruciating pain centred on the retro-orbital and orbital regions. These episodes lasted 1-4 h, recurring 2-3 times daily. The attacks were associated with ipsilateral ptosis, conjunctival injection, lacrimation, rhinorrhoea and facial flushing. From 1986 to 1988, he had trials of medications without any benefit. In February 1988, he had complete surgical section of the left trigeminal sensory root that shortened the attacks in length for 1 month without change in their frequency or character. In April 1988, he had further surgical exploration and the root was found to be completely excised; post-operatively, there was no change in the symptoms. From 1988 to 1999, he had a number of medications, including verapamil and indomethacin,;all of which were ineffective. Prednisolone 30 mg orally daily rendered the patient completely pain free. Sumatriptan 100 mg orally and 6 mg subcutaneously aborted the attack after similar to45 and 15 min, respectively. He was completely anaesthetic over the entire left trigeminal distribution. Left corneal reflex was absent. Motor function of the left trigeminal nerve was preserved. Neurological and physical examination was otherwise normal. MRI scan showed a marked reduction in the calibre of the left trigeminal nerve from the nerve root exit zone in the pons to Meckel's cave. An ECG-gated three-dimensional multi-slab MRI inflow angiogram was performed. No dilatation was observed in the left internal carotid artery during the cluster attack. Blink reflexes were elicited with a standard electrode and stimulus. Stimulation of the left supraorbital nerve produced neither ipsilateral nor contralateral blink reflex response. Stimulation of the right supraorbital nerve produced an ipsilateral response with a mean R2 onset latency of 36 ms and a contralateral response with a mean R2 onset latency of 32 ms. Lack of ipsilateral vessel dilatation makes the role of vascular factors in the initiation of cluster attacks questionable. With complete section of the left trigeminal sensory root the brain would perceive neither vasodilatation nor a peripheral neural inflammatory process; however, the patient continued to have an excellent response to sumatriptan. The case illustrates that cluster headache may be generated primarily from within the brain, and that triptans may have anti-headache effects through an entirely central mechanism.