Article Text


Cluster headache-like attack as an opening symptom of a unilateral infarction of the cervical cord: Persistent anaesthesia and dysaesthesia to cold stimuli
  1. Vincent de la Sayettea,
  2. Stéphane Schaeffera,
  3. Oguzhan Coskunb,
  4. François Leprouxb,
  5. Gilles Defera
  1. aService de Neurologie, bService de Neuroradiologie, CHU Côte de Nacre, Caen Cedex, France
  1. Dr V de la Sayette, Service de Neurologie, CHU côte de Nacre, F14033 Caen Cedex, France.


A 54 year old man experienced excruciating left retro-orbital pain with lacrimation and redness of the eye representative of a cluster headache attack. This was followed by left hemiparesis with plegia of the lower limb and left Horner’s syndrome. Five days later the hemiparesis recovered while the patient developed hypoanaesthesia to cold stimuli that evoked painful burning dysaesthesia on the right side below the C4 level. MRI disclosed a discrete infarct in the left lateral aspect of the cord at C2 level concomitant to a left vertebral artery thrombosis. This limited infarct and the clinical symptoms suggest a hypoperfusion in the peripheral arterial system of the left hemicord, supplied both by the anterior and posterior spinal arteries. Cluster headache-like attack and persistent dysaesthesia to cold stimuli are discussed respectively in view of the central sympathetic involvement and partial spinothalamic system dysfunction.

  • hemicord infarction
  • cluster headache attack
  • dysaesthesia
  • allodynia

Statistics from

The cervical cord is supplied by two arterial systems, central and peripheral, which overlap but are discrete. The first is dependent entirely on the single anterior spinal artery (ASA). The second, without clear cut boundaries, receives supplies from the ASA and both posterior spinal arteries.1 Because the ASA is medial and dominant, unilateral cord infarctions are very rare. They may occur in the perfusion territory supplied by the ASA2-5 as a result of the obstruction of either a duplicated ASA2 or the obstruction of one of the sulcal arteries which arise from the ASA and turn alternatively left or right, to supply one side of the central cord.3 Peripheral hemicord infarction may result from ischaemia in the territory of the ASA5 or posterior spinal artery.6 We report on a patient with a unilateral infarction limited to the left lateral column of the cervical cord in relation to a vertebral artery thrombosis. The clinical feature was peculiar in that the cluster headache-like attack presented as the opening symptom and, thereafter, the occurrence of anaesthesia and dysaesthesia to cold stimuli (allodynia) that affected the right side below C4 became apparent.

Case report

A 54 year old man with a history of mild hypertension (treated with 50 mg Iosartan/day) and common migraine without known autonomic feature presented on 13 October 1997 with an abrupt boring and throbbing pain in the left retro-orbital region spreading to the jaw, and the supraorbital, and temporal regions. There were photophobia, abundant lacrimation, and redness of the left eye. The excruciating pain lasted 30 minutes, then improved and disappeared about 20 hours later. Half an hour after onset of the headache, the patient experienced paraesthesia of the left lower limb which progressed, within 3 hours, to weakness of the left foot. The next days, a left hemiparesis occurred. The patient was admitted to our department on 15 October 1997.

His arterial pressure was initially 160/100 mm Hg and general examination showed no abnormalities. Neurological examination disclosed a left Horner’s syndrome and a flaccid left sided hemiparesis affecting the lower limb severely and partially the proximal part of the upper limb. The hand was slightly weak and the face was spared. Bilateral Babinski’s signs were present. Light touch, vibration, and sense of position were normal. Pinprick and cold and warm thermal stimuli were perfectly discriminated on both sides although felt appreciably more strongly on the left. There was equal facial sensation to light touch, pain, and temperature, including symmetric corneal response, despite the patient having reported persistent soreness and left hypersensitivity of the face and head to touch since the initial attack of retro-orbital pain. Five days after onset, the hemiparesis cleared almost completely while the patient experienced unpleasant and painful right sided provoked dysaesthesia. At that time, examination of the patient’s sensibility disclosed normal and symmetric sensations to light touch, vibration, sense of position, and pin prick. Temperature perception and discrimination were normal above C4. Below C4, on the right side, the thermal perception for warm stimuli (>28°C) was unchanged (quickly and strongly felt on the left side) and contrasted with more severe hypoaesthesia (upper limb) and anaesthesia (trunk and lower limb) to cold stimuli (<28°C). Moreover, cold stimuli (<28°C) evoked severe dysaesthesia described as burning, aching sensations radiating around the place of stimulation. Any stimulus other than cold could elicit them.

Normal investigations included routine laboratory examinations, CT of the head, transoesophageal echocardiography, and Holter monitoring. Brain MRI was unremarkable except for a left punctiform low void vertebral artery. MRI of the cervical spine, 11 days after the onset, disclosed a spontaneous abnormally increased signal on T2 and proton density weighted images in the lateral aspect of the left cord at C2 level (figure). Vertebral angiography showed a thrombosis of the left vertebral artery. A diagnosis of infarction in the left upper cervical hemicord was made. The patient was treated with heparin and subsequently fluindione. Four months later his motor strength fully recovered. A left Babinski’s sign and Horner’s syndrome persisted. The patient continued to complain of left soreness and hypersensitivity of the face and head and, above all, of a bitter, burning dysaesthesia to cold stimuli, refractory to carbamazepine, clonazepam, and amitriptyline.

MRI (GE-Signa 1.5 Tesla): (A) sagittal, (B) axial, and (C) coronal T2 weighted images (TR=5000 ms; TE=105 ms) of the cervical spinal cord 11 days after onset. Focal high intensity image in the left lateral aspect of the cord at C2. The lesion extends medially but both the anterior and posterior fasciculus are spared.


The combination of neurological signs, left flaccid hemiparesis sparing the face with plegia of the lower limb, Horner’s syndrome, and slight spinothalamic hypoaesthesia on the contralateral side, suggests a lesion located in the periphery of the left lateral column of the cervical cord. A T2 weighted cervical MRI disclosed a focal high intensity signal within the left lateral aspect of the cord, at C2 level, consistent with a limited infarct of the hemicord.4 7 Angiography documented further a left vertebral artery thrombosis that could correspond, although without certainty, to a dissection. Our patient shared the main clinical features found in the ASA hemicord syndrome2-5—namely, Horner’s syndrome and hemiparesis that progress over many hours on the side ipsilateral to the infarction while a spinothalamic hypoaesthesia develops on the contralateral side. None of these patients in the above mentioned studies had either posterior column loss or facial hypoaesthesia, both of which are relevant to posterior spinal artery syndrome.6 Even though our patient is similar in many aspects to others, two particular points have to be emphasised: (1) the initial excruciating retro-orbital pain; (2) the limited anaesthesia and persistent dysaesthesia to cold stimuli which developed soon after the onset.

All the reported cases of cervical hemicord infarction began with severe pain that involved predominantly the neck2-5 and often radiated to the shoulder or the upper limb. None of them presented with headache. On the contrary our patient never had cervicobrachial pain but developed extremely severe unilateral headache as an opening symptom. It could account for pain from arterial dissection8 but, in this outlook, rather from a carotid than a vertebral artery dissection.9 In return, the excruciating boring and throbbing pain in the left retro-orbital region, associated with autonomic features, lacrimation, and conjuctival injection, satisfied principally the criteria of cluster headache attack.8 10 Nevertheless, as our patient experienced only a single episode of retro-orbital pain which constrasted with the striking rhythmicity of typical cluster headache attack, the generic terminology of trigeminal-autonomic cephalagia (TAC) seems more appropriate.11 This kind of intense short lasting headache with autonomic features including cluster headache attack, chronic paroxysmal hemicrania, episodic paroxysmal hemicrania, and SUNCT syndrome (short lasting unilateral neuralgiform headache with conjunctival infection and tearing), shares a linked pathophysiology.11 Trigeminal-autonomic cephalagia is likely to be the consequence of an activation of the trigeminal nerve and cranial parasympathetic nervous systems11-13 through the trigeminal-autonomic reflex.11 The finding presented here suggests a central mechanism for the initiation of attack although the causal relation with the lesion needs to be interpreted with caution. A dysfunction of the trigeminal system is doubtful as there was no clinical evidence of trigeminal nerve involvement whence the spinal nucleus is supplied by branches of the posterior spinal artery. Interestingly, the persistent residual soreness and hypersensitivity of the head and face to touch has been already reported in some patients after a cluster headache attack.10 It might be postulated that the lesion of the central sympathetic pathway in the left cervical cord was the primary mechanism which allowed the attack to develop; as a consequence, the patient retained a residual Horner’s syndrome. Indeed, there is abundant evidence of sympathetic dysfunction, whether peripheral or central, during the cluster period of headache.12 14-16 The first phase of attack may include impaired sympathetic nervous activity which should induce increased susceptibility to attack provocation.12 Symptomatic attacks of cluster headache have been described in a wide variety of disorders, most of them located in the middle fossa close to the midline and the cavernous sinus region17 or, more rarely, outside or inside the cervical spine.10 Three cases might be considered as their lesions were located laterally near the medulla and upper cervical cord. They concern an upper cervical menigioma,18 a vertebral artery aneurysm,19and a vertebral artery dissection.20 The common mechanism for triggering attack in the above mentioned cases might be the trigeminal-autonomic activation from upper cervical segments which overlap and connect with the spinal trigeminal nucleus,18 20 a hypothesis that should be considered in our own case.

The last particular point was the occurrence of painful dysaesthesia. This kind of burning stinging pain, which usually, accompanies dysfunctions of the spinothalamic pathway,21-23 is very rare after cord lesions24 especially spinal cord infarcts.21 It can be provoked by non-noxious stimuli in the deafferented dermatomes (allodynia), either by thermal stimuli23 or, in most cases, gentle touching. Therefore it is usually postulated that dysaesthesia might result from an afferent imbalance between an impaired spinothalamic system and an intact lemniscal sensory input.22 Dysaesthesia in our patient could hardly correspond to this kind of mechanism as they were evoked only by cold stimuli. There was clinical evidence of partial involvement of the spinothalamic system as touch and pinprick were normally felt contrasting with hypaesthesia to thermal stimuli and above all anaesthesia (trunk and lower limb) to cold stimuli. Thus considering dysaesthesia in our patient in terms of imbalance within the spinothalamic system itself seems an attractive hypothesis. Interestingly dysasthesia occurred while the patient recovered from hemiparesis and hypoaesthesia to pinprick and developed anaesthesia to cold stimuli. This response (dysaesthesia) suggests a functional rearrangement so that the cold stimuli could raise an inappropriate activation of pain message instead of cold thermic sensation. In this way, Triggs and Beric22 considered such an alternative explanation as did Nathan and Smith,24 who suggested that dysaesthesia may result from selective lesions of the spinothalamic tract sparing the spinoreticulothalamic connections.

Considering the anatomical organisation of the two spinothalamic pathways (anterior and posterior tracts) and the topographical arrangement of the fibres that they convey25 (crude tactile, pain, and thermal sensations), clinical symptoms suggest, in our patient, lesioning in the most dorsal and superficial part of the spinothalamic tract. This accords well with the initial hemiparesis with plegia of the lower limb which, in turn, also indicate (considering the somatotopic arrangement of the fibres within the corticospinal tract) a superficial lesion of the lateral cord. MRI confirmed the small size of the lesion, which was confined to the lateral fasiculus and limited in height. The deep extension could correspond to oedema7 rather than to ischaemia. This suggests a hypoperfusion in a “watershed area” between both anterior and posterior spinal peripheral arterial systems rather than ischaemia in the perfusion territory of a sulcal artery or a duplicated ASA. This hypothesis agrees with the common origin of the anterior and posterior spinal arteries which both arise from the vertebral artery.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.