Cluster a brief severe associated with autonomic signs and symptoms A - PDF document

Cluster a brief severe associated with autonomic signs and symptoms (Appendix 1).The signature feature of this fascinating headache disorder is itsrhythmicity. Cluster headache is unique in that it displays both acircadian and circannual periodicity. Cluster headache periodstend to recur at the same time of year for a given patient and atthe same time of day, with clock-like regularity.Though clusterheadache has thought of as a vascularheadache disorder, the periodicity of cluster headache suggestsinvolvement of the hypothalamus and, more specifically, thesuprachiasmatic nucleus of the hypothalamus, the biologicalclock. This review will discuss the evidence for a disorder ofcircadian rhythm and hypothalamic function in cluster headache.ABSTRACT:This article reviews the literature for evidence of a disorder of circadian rhythm andhypothalamic function in cluster headache. Cluster headache exhibits diurnal and seasonal rhythmicity.While cluster headache has traditionally been thought of as a vascular headache disorder, its periodicitysuggests involvement of the suprachiasmatic nucleus of the hypothalamus, the biological clock. Normalcircadian function and seasonal changes occurring in the suprachiasmatic nucleus and pineal gland arecorrelated to the clinical features and abnormalities of circadian rhythm seen in cluster headache.Abnormalities in the secretion of melatonin and cortisol in patients with cluster headache, neuroimagingof cluster headache attacks, and the use of melatonin as preventative therapy in cluster headache arediscussed in this review.While the majority of studies exploring the relationship between circadianrhythms and cluster headache are not new, we have entered a new diagnostic and therapeutic era inprimary headache disorders. The time has come to use the evidence for a disorder of circadian rhythmin cluster headache to further development of chronobiotics in the treatment of this disorder.RƒSUMƒ: CŽphalŽe vasculaire de Horton :donnŽes en faveurdÕun dŽsordre du rythme circadien et de lafonction hypothalamique.Cet article revoit la littŽrature sur les donnŽes en faveur dÕun dŽsordre du rythmecircadien et de la fonction hypothalamique dans la cŽphalŽe vasculaire de Horton. Celle-ci a une rythmicitŽ diurneet saisonnire. Bien quÕelle Žtait traditionnellement considŽrŽe comme une cŽphalŽe vasculaire, sa pŽriodicitŽsuggre lÕimplication du noyau suprachiasmatique de lÕhypothalamus, lÕhorloge biologique. La fonction circadiennenormale et les changements saisonniers survenant dans le noyau suprachiasmatique et la glande pinŽale sont corrŽlŽsaux manifestations cliniques et aux anomalies du rythme circadien observŽes dans les cŽphalŽŽ de Horton. Desanomalies de la sŽcrŽtion de la mŽlatonine et du cortisol chez ces patients, la neuroimagerie des accs de cŽphalŽeet lÕutilisation de la mŽlatonine comme traitement prŽventif des cŽphalŽes sont discutŽes dans cette revue. Bien quela majoritŽ des Žtudes sur la relation entre le rythme circadien et la cŽphalŽe de Horton ne soient pas rŽcentes, noussommes entrŽs dans une nouvelle re diagnostique et thŽrapeutique dans le domaine des cŽphalŽes primaires. Il esttemps dÕutiliser les donnŽes en faveur dÕun dŽsordre du rythme circadien dans la cŽphalŽe de Horton pourpromouvoir le dŽveloppement de mŽdicaments chronobiotiques dans le traitement de cette maladie. Can. J. Neurol. Sci. 2002; 29: 33-40 THE CANADIAN JOURNALOF NEUROLOGICALSCIENCES33 ClusterHeadache: Evidence foraDisorderof Circadian Rhythm andHypothalamic Function Tamara Pringsheim* * Dr Pringsheim is the recipient of the first Canadian Headache Society JacquesMeloche Annual Headache Prize, awarded to a Canadian Neurology Resident for thebest essay on a headache-related topic.From the Department of Neurology, University of Toronto, Toronto, ON Canada.R ECEIVED J UNE 25, 2001. A CCEPTED O 5, 2001.Reprint requests to:Tamara Pringsheim, Resident in Neurology, University of Toronto,c/o Dr WJ Becker, Division of Neurology, 12th Floor, Neurology, Foothills Hospital,1403-29th Street NW, Calgary AB T2N 2T9 Canada. E-mail: tpringsh@uhnres.utoronto.caThe normal biology of circadian rhythms and function will besummarized, followed by a discussion of the abnormalities ofcircadian function seen in cluster headache. While the majorityof the seminal research done in this field occurred more than KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV twenty years ago, the topic of circadian rhythms in clusterheadache is worth revisiting. This review attempts to integratethe clinical features and abnormalities of circadian rhythm seenin cluster headache with functional knowledge of the circadiantiming system. M OLECULAR C IRCADIAN B IOLOGY The circadian rhythm of behaviours and hormones arisesfrom a rhythm at the level of expression of clock genes. The firstclock gene was discovered in mutated Drosophila melanogasterfruit flies dis

playing unusual circadian rhythms and was mappedto the period (per) locus on the X chromosome. 3 More recently,the timeless (tim) gene was identified. 4 The protein products of per(PER) and tim(TIM) are believedto travel between the cytoplasm and nucleus of cells, andregulate their own expression, as well as that of target genes(Figure 1) 5 . PER and TIM bind together to function. Both genesare transcribed in the morning and their mRNAs accumulate inthe daytime. The TIM protein is degraded by light, which causesa low level of TIM protein despite tim RNAtranscription. PERcannot function without TIM, therefore, during the day theprotein products are non-functional. After dusk, TIM levels areable to increase, and the two proteins can bind together andbecome functional. They enter the cell nucleus and inhibit theirown genes and target genes. Subsequently,perand tim mRNAlevels decrease and protein expression decreases. By morning,PER and TIM protein levels are low and transcription is nolonger inhibited Ð the cycle begins again. It is believed that asimilar mechanism in mammals drives the circadian timingsystem. 6 Clock genes provide the biological mechanism thatgenerate circadian rhythms.T HE S UPRACHIASMATIC N UCLEUSOFTHE H YPOTHALAMUS The hypothalamus is the principle central regulator andintegrator of the endocrine system. The synthesis and release ofhormones from the anterior lobe of the pituitary gland isregulated by peptides secreted from the hypothalamus into thehypothalamo-hypophyseal portal system. The hypothalamuscontrols the output of hormones that have important effects onthe modulation of sexual function and behaviour, thyroidsecretion, cortisol secretion, stress responses, appetite, growth,temperature, and salt T h ehypothalamus lies beneath the thalamus, and can be divided intothree main areas with distinct groups of nuclei: the rostralsupraoptic, middle tuberal and posterior mamillary areas. Thesuprachiasmatic nucleus of the hypothalamus is located in therostral supraoptic area.In is thesuprachiasmatic nucleus. Destruction of the suprachiasmaticnucleus induces the disappearance of circadian rhythmicity ofhormone release 7 and sleep-wakefulness cycles. 8 Transplantationof neonatal rat suprachiasmatic nuclei into arrhythmic ratsrestores 9 endogenously generated but entrained to daily light-dark cyclesof the external environment by the retinohypothalamic tract. Thecircadian timing system consists of three major components: THE CANADIAN JOURNALOF NEUROLOGICALSCIENCES34 Figure 1:PER/TIM interaction. KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV photoreceptors from visual pathways mediating entrainment; apacemaker of circadian rhythm Ð the suprachiasmatic nucleusitself; and output from the pacemaker that produces an entrained 10 In humans, the suprachiasmatic nucleus is a paired group ofcells lying dorsal to the optic chiasm and lateral to the thirdventricle. The suprachiasmatic nucleus in humans is not as welldifferentiated as in other mammals, and there is variability instructure between subjects. The shape of the nucleus is sexuallydimorphic, appearing elongated in women, and more spherical inmen. 11 Distinct populations of vasopressin, vasoactive intestinalpeptide (VIP), neuropeptide Y, and neurotensin neurons havebeen the T h eretinohypothalamic pathway terminates in a distinct subdivisionof the suprachiasmatic nucleus, characterised by the presence ofVIPneurons. 10 Amarked seasonal variation has been observed in thevolume, total cell number and number of vasopressin cells of thehuman 1 2 The volume of thesuprachiasmatic nucleus is twice as large in the autumn as in thesummer and contains almost twice as many cells. Similarseasonal in the number of containing neurons. The suprachiasmatic nucleus is smaller insummer than in any other season. In contrast to the annualvariations in the human suprachiasmatic nucleus, no diurnalvariations have been observed. However, in rats, neuronal inputsfrom V I Pimmunoreactivity and VIPmRNAlevels during the daytime. 13 VIPand VIPmRNAreach a peak level at 02h00 and a trough at14h00. It appears that the expression of VIPand its mRNAin theVIPneurons of the suprachiasmatic nucleus is light-dependent.The annual variation in the total cell number of the humansuprachiasmatic nucleus has been correlated to the photoperiodiccycle of the temperate zone. Peaks in the total cell number of thesuprachiasmatic nucleus occur close to the spring and autumnalequinox. The annual minimum of total cell number was found tocoincide with the summer solstice, with a second smallerminimum occurring at the winter solstice. These findings showthat the suprachiasmatic nucleus undergoes its greatest increasein size when the photoperiod undergoes its largest rate of change.The shortening days of autumn, as well as the lengthening daysof spring, appear to induce morphological changes in thesuprachiasmatic nucleus. 14 Circadian function and behaviourparallel dynamic changes within the suprachiasmatic nucleus.M ELATONINANDTHE P INEAL G LAND The hormone melatonin is synthesised in the pineal gland.There is a daily rhythm of melatonin production, with peaklevels occurring in the

hours of darkness. The melatonin rhythmis endogenously generated by the suprachiasmatic nucleus.Environmental light entrains melatonin secretion to a 24-hourcycle.that pass retinohypothalamic tract to the suprachiasmatic nucleus, then tothe superior cervical ganglion, and finally to the pineal gland.The release of norepinephrine to the pineal gland stimulatesmelatonin photoreceptor cells are hyperpolarized, which inhibits the releaseof 1 5 Further the pathway, gama-aminobutyric acid (GABA) release from the suprachiasmaticnucleus to the sympathetic system mediates the inhibitory effectof light on melatonin release by the pineal gland. 1 6 Wi t hdarkness, the photoreceptors release norepinephrine, and thesystem is activated. 15 The pineal gland displays seasonal variations, being larger inwinter than summer in volume and in the nuclear size ofp i n e a l o c y t e s . 1 7 The pattern of melatonin synthesis is alsoinfluenced by the photoperiod. During long photoperiods, suchas the summer, the period of high melatonin synthesis isshortened, in contrast to the prolonged synthesis of melatoninduring wintertime. 18 Melatonin feeds back onto the suprachiasmatic nucleus.Melatonin receptors been found on neurons, providing the link for a suprachiasmaticnucleus-pineal gland feedback loop. Sack et al 19 propose that thisfeedback loop mediates circadian phase shifting and perhapsattenuates Melatonin shifts the circadian clock. Daily injections ofmelatonin produced phase advances sufficient to entrain a 24hour cycle in rats that were previously free running with acircadian period greater than 24 hours, in constant dim lightc o n d i t i o n s . 2 0 The melatonin have been applied to several disorders, such as jet lag,delayed sleep phase syndrome, and non-24-hour sleep-wakedisorder.The phase shifting effects of melatonin are dependent on thetiming of melatonin administration and can occur with nearphysiologic doses. Administration of melatonin in the afternoonand early evening will advance (shift earlier) the phase of thecircadian rhythms. Administration in the morning at 07h00 willdelay (shift later) the phase. 2 1 Knowledge of the patientÕscircadian phase prior to treatment will allow accurate timing ofthe melatonin dose in order to achieve the desired phase shift.There is a correlation between circulating melatonin levelsand sleep propensity. Evidence suggests that along with itschronobiotic have some sleeppromoting activity. Daytime administration of melatonin whenthe endogenous levels of the hormone are low was shown toinduce subjective feelings of sleepiness and fatigue, and improvesleep quality. 2 2 The hypnotic effect of melatonin may beindependent of its synchronising influence on the circadianrhythm and may be mediated by a lowering of core bodytemperature. The evening rise of melatonin is temporally relatedwith the of core body pharmacologic doses of melatonin can induce a decrease in bodytemperature. 21 Melatonin functions as a chronobiotic, capable ofmodifying and adjusting the circadian timing system.C ORTISOLAND H YPOTHALAMIC P ITUITARY A XIS Cortisol release from the adrenal cortex also exhibits a dailyrhythm. The anterior pituitary corticotroph cells synthesizeadrenocorticotropic hormone (ACTH), which is released afterstimulation of the cells by corticotropin-releasing hormone LE JOURNALCANADIEN DES SCIENCES NEUROLOGIQUESVolume 29, No. 1 Ð February 200235 KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV (CRH). Corticotropin-releasing hormone release is controlled byseveral neurotransmitter systems, including acetylcholine andserotonin, GABAa n dnorepinephrine, which inhibit its release. Stress related stimuliwill also affect CRH release. Release of both CRH and ACTH ismodulated through a negative feedback effect of cortisol.Selective lesions of the suprachiasmatic nucleus in the rat lead toa loss of the rhythmic secretion of cortisol, with cortisol levels atall time points remaining in the range normally observed formorning values. 7 Cortisol diffuses into its target cells, combines with receptorproteins in the cytoplasm, and is transported to the nucleus whereit modifies gene expression and protein synthesis. Its manyeffects on metabolism include stimulation of hepatic glucosesynthesis, breakdown of protein in muscle, skin and bone, ananti-inflammatory effect, and enhancement of lipid mobilisation.Cortisol protects against stress-induced hypotension, shock anddeath by mechanisms that are not well understood. 23 The cortisolrhythm is an example of how the circadian timing systemcontrols physiology to enhance adaptation and survival.A BNORMALITIESOF C IRCADIAN R HYTHMIN C LUSTER H EADACHE It has long been observed that cluster headache periods tendto recur at the same time of year for patients. This led Kudrow 24 to a retrospective study in 404 patients with episodic clusterheadache, noting the mean monthly frequency of cluster periodonset, with consideration of the mean monthly duration ofdaylight. The frequency of cluster period onset increased withthe gradual increase or decrease in daylight throughout the year,with peaks occurring seven to 10 days after the longest andshortest days of t

he year. Each peak lasted the months of Januaryand July.The gradual rise of cluster period frequency wasinterrupted a in beginning seven to 10 days after the resetting of clocks forDaylight Savings time in April and Standard time in October.The author hypothesised that the cluster period may somehowresult from an inability to synchronise the internal circannualpacemaker to environmental light clues. 24 Cluster headache exhibits a relationship with the sleep-wakecycle and with the activity-relaxation cycle. Attacks may occurat the same hour every day. 25-27 Cluster headaches are known tooccur during sleep in up to two-thirds of patients. 28 Manzoni etal 29 found that when plotting the most common hours of onset forall patients, sharp peaks were found between 01h00 and 02h00,13h00 and 15h00, and a final peak reached at around 21h00. Thisstudy was performed in Italy and the majority of the patientsstopped working between 13h00 and 15h00. Russell 30 found that71% of daytime attacks in his series occurred while patients werephysically relaxed. Nocturnal attacks usually begin one to two hours after fallingasleep and, in patients with episodic cluster headache, areassociated with rapid eye movement sleep. 31 Almost 60% ofrecorded attacks in one series followed REM sleep, while REMcomprised only 20% of total sleep time. 32 Patients are able tohave a transient remission of their cluster headache by skippingone nights sleep. 33 Abnormal sleep has also been reported in patients during thecluster period. Subjective assessment of cluster patients incomparison to healthy controls has shown a reduced totalduration of nocturnal sleep, increased time to fall asleep, a highfrequency of nocturnal awakenings, and poor sleep quality.Objective polysomnographic findings were also significantlydifferent in cluster patients than the control group, with clusterpatients exhibiting a reduced total duration of sleep, increasedtime to fall asleep, increased latent periods of all sleep stages,increased percentage representation of superficial sleep stages(stages 1 and 2), reduced duration of REM stage, increasednumber of awakening episodes during sleep, and increasedmovement activity. 34 Sleep apnea is a common finding inpatients with cluster headache and nocturnal attacks have beenfound in association with oxyhemoglobin desaturation. 32 Thesestudies indicate that not only do cluster attacks occur duringsleep but sleep patterns themselves are altered in patients withthis disorder.A LTERATIONSINTHE S ECRETIONOF M ELATONINAND C ORTISOLIN C LUSTER H EADACHE Alterations in the circadian secretion of hormones has alsoprovided evidence of deranged hypothalamic function in clusterheadache. The first evidence of hypothalamic involvement incluster headache came from the demonstration of loweredthe headache period in men. 35 While this review will focus on theabnormalities in the secretion of melatonin andcortisol, alterations in the secretion of luteinizing hormone andprolactin, and altered responses of luteinizing hormone, folliclestimulating hormone, prolactin, growth hormone, and thyroidstimulating hormone to challenge tests have been demonstratedin patients with cluster headache. 36 One of the first studies of melatonin and cortisol secretion inepisodic cluster headache was performed by Chazot et al. 37 Melatonin and cortisol secretion was studied over a 24-hourperiod, with blood collected every two hours during the day andat one hour intervals at night, in healthy controls and patients intheir cluster bout. Most cluster patients had lowered nocturnalmelatonin secretion, particularly in the early part of the night.Cluster patients displayed a significant phase advance in theirmelatonin rhythm, with the acrophase (time from midnight topeak hormone levels) occurring two hours earlier than controlsubjects. For cortisol, the rhythm appeared slightly blunted in thecluster headache group and was significantly phase advanced.The authors concluded that their data are compatible with adisturbance located in the of thehypothalamus, though the blunting of the cortisol rhythm couldbe due to a number of other causes as well. They emphasized thepathophysiological of the sympathetic superiorcervical ganglion, which provides sympathetic innervation to thepineal gland, cranial blood vessels, choroid plexus, the eye,carotid body, and salivary and thyroid glands. The authorshypothesised that the unilaterality of cluster attacks and theassociated autonomic features may be related to abnormalmodulation from the superior cervical ganglion level, withimpaired secretion a thisabnormality. 37 Waldenlind et al 38 studied 24-hour rhythms of serum cortisol THE CANADIAN JOURNALOF NEUROLOGICALSCIENCES36 KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV and melatonin in patients with cluster headache during andbetween cluster periods and in healthy controls. Compared tohealthy controls, lower maximal nocturnal serum melatoninlevels were found in cluster headache, both in the active periodand in remission. Maximal cortisol levels and 24-hour meanswere significantly higher during the active cluster period but notduring clinical remission, as compared to healthy co

ntrols. Adelay of the cortisol minimum was found in the active clusterperiod as compared to the The authorshypothesised that while the phase delay of cortisol may reflect animpaired chrono-organization of hormone secretion in patientswith cluster headache, the higher 24-hour secretion of cortisolmight represent an adaptive response to the pain of clusterattacks. 38 Leone et al 39 also studied the circadian secretion of melatoninand cortisol in patients with episodic cluster headache, during thecluster period, as compared to healthy controls. They foundsignificantly lower plasma melatonin levels in cluster headachepatients than controls, and no significant melatonin rhythm inone-third of their subjects. Cortisol 24-hour production wassignificantly higher in cluster headache patients than controls,while the amplitude and acrophase were similar.Almost half oftheir subjects with cluster headache had no cortisol rhythm. Inaddition, in healthy subjects, the timing of the melatoninthat of was nothe authorsconcluded that the absence of any correlation between the painparameters and the circadian production of these hormonessuggested a primary derangement in the biological clock duringthe cluster period. 39 The reduction in night-time melatoninsecretion and loss of the melatonin rhythm reported in thesestudies may reflect dysfunction within the synthetic pathway ofmelatonin production with consequent loss or alteration of itscircadian phase shifting properties.Further studies have assessed the cortisol and ACTH responseto the insulin tolerance test and ovine CRH test. The insulintolerance test produces hypoglycemia, induces a stress response,which increases CRH release and therefore ACTH and cortisolsecretion. This test measures the integrity of the hypothalamic-pituitary-adrenal (HPA) axis and its ability to respond to stress.Ovine CRH stimulation assesses ACTH secretory dynamics.This test is used to diagnose primary and secondary adrenalinsufficiency. 40 Leone et al 41 found that both remission andcluster period patients had significantly higher basal cortisollevels than controls. Ablunted cortisol response to ovine CRHwas found in both active cluster and the remission phase, thoughthe ACTH surge was normal. Areduced cortisol and ACTHresponse was found to the insulin tolerance test in both phases ofthe disorder. 41 The authors repeated the study in a group ofpatients with low back pain due to disc herniation and did notfind similar responses to these tests. They concluded that thealtered HPAaxis responsiveness in cluster headache patients isnot a consequence of pain but rather due to hypothalamicderangement. 42 The abnormal results of the insulin tolerance testpoint to dysfunction along the HPAaxis in patients with clusterheadache, while the results of the ovine-CRH test suggest anabnormal adrenal response. Anormal ovine-CRH test wouldhave more theis thesuprachiasmatic nucleus is involved in the setting of sensitivityof the adrenal cortex to ACTH, with experimental demonstrationof a pathway. 43 Therefore, these findings could be consistent withhypothalamic dysfunction. Areduced response of the HPAaxisto the insulin tolerance test has been reported in patients withbrain tumours and multiple system atrophy involving theh y p o t h a l a m u s . 4 2 One possible explanation for the changesobserved in cortisol rhythm, secretion, and response to challengetests may be failure of the circadian timing system in theentrained rhythm of physiologic processes.M ELATONINAS P REVENTATIVE T HERAPYIN C LUSTER H EADACHE Leone et al 4 4 performed a study ofmelatonin versus placebo in the prophylaxis of cluster headache.Twenty patients with cluster headache (18 episodic, two chronic)participated in the study. Patients with episodic cluster headacheentered the study between the second and tenth day of theircluster bout. After a run-in period of one week withoutprophylactic treatment, patients were randomized to receive 10mg melatonin or placebo for two weeks. The authors found thatcompared to the run-in period, there was a reduction in the meannumber of daily attacks and a strong trend towards reducedanalgesic consumption in the melatonin group but not in theplacebo group. Five patients in the melatonin group responded tothe treatment, with cessation of cluster headaches after five daysof treatment. 44 N EUROIMAGINGOF C LUSTER H EADACHE More recently, positron emission tomography (PET) has beenused to assess changes in regional cerebral blood flow as anindex of synaptic activity during nitroglycerin-induced clusterattacks. May et al 45 found that in the acute pain state, activationwas seen in the ipsilateral inferior hypothalamic grey matter, thecontralateral ventroposterior thalamus, the anterior cingulatecortex and bilaterally in the insula, and in the cerebellarhemispheres (Figures 2 and 3). The area of hypothalamicactivation occurred in the region of the circadian pacemakerneurons, further establishing the involvement of this area of thehypothalamus in the genesis of acute cluster attacks. 45 APETstudy of patients in and out of their cluster bout demonstratedposterior hypothalamic grey activation only in patients withnitroglycerin-induced cluster headache attacks in the activecluster period. Hypothalamic grey activation was not seen innitroglyc

erin-induced headaches out of the bout. Activation ofthe hypothalamus therefore appears to be specific to clusterh e a d a c h e . 4 6 In morphometry T 1 -weighted MRI scans has revealed an increase in hypothalamicvolume, located in the inferior posterior hypothalamus in clusterheadache patients compared to normal controls. 47 D ISCUSSIONAND C ONCLUDING R EMARKS Several correlations can be made between the biology ofcircadian rhythms and hypothalamic function and the clinicalfeatures of cluster headache. It is interesting to observe that boththe suprachiasmatic nucleus and pineal gland undergo a change LE JOURNALCANADIEN DES SCIENCES NEUROLOGIQUESVolume 29, No. 1 Ð February 200237 KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV in size circannually, due to the lengthening and shortening ofdaylight hours. The suprachiasmatic nucleus undergoes itsgreatest change in size when the photoperiod undergoes itslargest rate of change, just as the incidence of cluster boutsincreases with the gradual increase or decrease in daylightthroughout the year. One may hypothesise that perhaps a failureor dysfunction within the suprachiasmatic nucleus occurs andmay be implicated in the pathogenesis of cluster headache.Adiurnal expression of VIPhas been demonstrated in the rat,with peak levels occurring at 02h00 and a trough at 14h00. If thisdiurnal expression of VIPoccurs in humans, the peak hours forcluster attacks (between 01h00 and 03h00) coincide with thispeak of VIPsecretion. VIPis a potent vasodilator, and elevationof VIPhas been reported in a study of the external jugular venousblood of cluster headache patients during a spontaneous clusterattack. 48 The investigators of this study concluded that theelevation of VIP, a marker of parasympathetic activity, was dueto activation of a brain stem reflex, the afferent arc of which isthe trigeminal nerve and the efferent the cranial parasympatheticoutflow from the seventh nerve. Perhaps a central derangementin the hypothalamic V I Psynthesising neurons can be anadditional explanation of these findings. If cluster headache is due to a problem in the hypothalamiccentre responsible for the synchrony of circadian rhythms, it islikely the phase resetting properties of melatonin that explain itsefficacy as a preventative treatment in some patients with clusterheadache. Appropriately timed melatonin therapy can shiftcircadian rhythms. Kudrow 24 hypothesised that the circannualincidence of cluster bouts may result from an inability tosynchronise the internal circannual pacemaker to environmentallight cues. Melatonin feedback onto the suprachiasmatic nucleusthrough melatonin receptors mediates this phase shifting effect.Perhaps it is the chronobiotic effect of melatonin on clusterheadache patients which is therapeutic. If this hypothesis is true,one would expect to see an effect of melatonin in clusterheadache using appropriately timed, physiologic doses of thecompound. Sack et al 19 have advanced this theory in theiranalysis of the sleep-promoting effects of physiologic doses ofmelatonin, which they believe are due to its phase-shiftingproperties. As of yet, only a supraphysiologic dose of melatoninhas been assessed as preventative therapy. If no response occursat physiologic doses, it may be the other properties of melatonin-potentiation GABA 4 9 inhibition 4 4 vasoconstriction 50 its mild hypnotic effect and ability to lower core bodytemperature, that explain its effect in cluster headache.Nagtegaal et al 51 reported a patient treated with melatonin fordelayed sleep phase syndrome who also had cluster headache.Both disorders responded to melatonin therapy. One mightexpect to find an increased incidence of circadian sleep disordersin patients with cluster headache, if a circadian influence isimplicated in its pathophysiology.While small series have THE CANADIAN JOURNALOF NEUROLOGICALSCIENCES38Figures 2 and 3: Comparison of nitroglycerin - induced acute cluster headache attack and rest (no pain) in nine patients with active chronic clusterheadache. Reprinted with permission from Elsevier Science (The Lancet, 1998; 352:275-78) Activations during the attack are shown as statistical parametric maps that show the areas of significant cerebral blood flow increases (p)colour superimposed on an anatomical reference derived from a T1 - weighted MRI. Significant activation was detected next to the third ventricleslightly lateralized to the left and rostral to the aqueduct. The activation is ipsilateral to the pain side and localized to the hypothalamic grey matter. KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV reported sleep dysfunction in patients with cluster headache, nostudies have assessed the prevalence of specific sleep disordersin this patient population.The effect of bright light therapy may be worth assessing inpatients with cluster headache. Light, like melatonin, resets thecircadian timing system according to a phase response curve. Inthe first half of the night, light exposure causes phase delays, andin the second half of the night, it c

auses phase advances. 52 Lighttherapy has been used in the treatment of circadian sleepdisorders, due to its ability to shift circadian rhythm, and inseasonal affective disorder, in which patients have a delay in theonset of melatonin secretion. If melatoninÕs effect is due to itsphase shifting properties, light therapy may have a similar effect.The majority of studies exploring the relationship betweencircadian rhythms and cluster headache are not new, in fact, themajority of articles discussing this relationship were published inthe late 70s, 80s and early 90s. However, we have entered a newdiagnostic and therapeutic era in primary headache disorders,with functional MRI and PETstudies demonstrating abnor-malities during both cluster 45-47 and migraine 53,54 attacks, andnew, highly effective, specific headache medications. With theaid of these sophisticated imaging techniques and the ability todevelop specific treatment for headache, the time has finallyarrived to put all the research done by our predecessors topractical application and for further development in the area ofchronobiotics in cluster headache.A CKNOWLEDGEMENTS The author thanks Drerner Becker for his encouragement and helpin the editing of the manuscript. At the time of the writing of thismanuscript, the author was the Canadian Headache Society clinical andresearch fellow, supported by a grant funded by the Canadian Institute ofHealth Research and GlaxoSmithKline. The author also thanks Drs. A r n eMay and Peter Goadsby for the use of their work in Figures 2 and 3.R EFERENCES 1.Headache Classification Committee of the International HeadacheS o c i e t y. Classification and diagnostic criteria for headachefacial pain. 2.Dodick DW, Rozen TD, Goadsby PJ, Silberstein SD. Clusterheadache. Cephalalgia 2000; 20:787-803.3.Konopka RJ, Benzer S. Clock D ro s o p h i l i amelanogaster. Proc Nat Acad Sci USA1971;68: 2112-2116.4. Barinaga M. New clock gene cloned. Science 1995;270:732-733.5.Gilliam TC, Kandel ER, Jessell TM. Genes and Behavior. In:Kandel ER, Schwartz JH, Jessell TM. Principles of NeuralScience, 4th Edition : McGraw-Hill 2000: 36-616.Myers MP,Wager-Smith K, Rothenfluh-Hilfiker A, Young MW.Light-induced degradation of TIMELESS and Entrainment of theDrosophila Circadian Clock. Science 1996;271:1736-1740. 7.Moore RY, Eichler VB. Loss of a circadian adrenal corticosteronerhythm following suprachiasmatic lesion in the rat. Brain Res1972; 42:201-206.8.Ibuka N, Kawamura H. Loss of circadian sleep-wakefulness cyclein the rat by nucleus Brain Res1975;96:76-81.9.Sawaki Y, Nihonmatsu I, Kawamura H. Transplantation of theneonatal nucleus rats with bilateral suprachiasmatic lesion. Neurosci Res 1984:1;67-72.10.Moore RIn: Swaab DF, Hofman MA, Mirmiran M, Ravid R, van LeeuwenF W. eds. Progress in Brain Research, vol 93. A m s t e r d a m :Elsevier. 1992:101-117.11.Swaab DF, Gooren LJG, Hofman MA. The human hypothalamus inrelation to gender and sexual orientation. In: Swaab DF, HofmanMA, Mirmiran M, Ravid R, van Leeuwen FWeds.. Progress inBrain Research, vol 93. Amsterdam: Elsevier. 1992:205-219.12.Hofman MA, Purba JS, Swaab DF. Annual variation in thevasopressin neuron population of the human suprachiasmaticnucleus. Neuroscience 1993;53(4):1103-1112. 13.Ibata Y, Okamura H, Tanaka M, et al. Functional morphology of thesuprachiasmatic nucleus. Frontiers Neuroendocrin 1999;20:241-268.14.Hofman MA, Swaab DF.The human hypothalamus: comparativemorphometry and photoperiodic influences. In: Swaab DF,Hofman MA, Mirmiran M, Ravid R, van Leeuwen FW, eds.Progress in Brain Research, vol 93. Amsterdam: Elsevier.1992:133-149.15.Brzezinski A. Melatonin in humans. N Eng J Med 1997;336(3):186-195.16. Kalsbeek A, Cutrera RA, Van Heerikhuize JJ, Vander Vliet J, BuijsR. GABArelease from SCN terminals is necessary for the lightinduced inhibition of nocturnal melatonin release in the rat.Neuroscience 1999;91(2):453-461.17.McNulty JA, Fox LM, Spurrier WA. Acircannual cycle inpinealocyte synaptic ribbons in the hibernating and seasonalreproductive 1990;119:237-240.18.Vanecek J, Illnerova H. Effect of short and long photoperiods onpineal N-acetyl transferase rhythm and on growth of testes andbrown adipose tissue in developing rats. Neuroendocrinology1985;41:186-191.19.Sack RL, Hughes RJ, Edgar DM, Lewy A. Sleep-promoting effectsof melatonin: at what dose, in whom, under what conditions, andby what mechanisms? Sleep 1997;20(10):908-915.20.Redman J, Armstrong S, Ng KT. Free-running activity rhythms inthe rat: entrainment by melatonin. Science 1983;219:1089-1091.21.Avery D, Lenz M, Landis C. Guidelines for prescribing melatonin.Ann Med 1998;30:122-130.22.Penev PD, Zee PC. Melatonin: a clinical perspective. Ann Neurol1997;42:545-553.23.Levy A, Lightman S. Endocrinology. Oxford University Press,1997.24.Kudrow L. The cyclic relationship of natural illumination to clusterperiod frequency. Cephalalgia 1987;7:76-78.25.Horton BT.The use of histamine in the treatment of specific typesof headache. JAMA1941;116:377-383.26.Ekbom KA. Ergotamine tartrate orally in HortonÕs ÒhistaminiccephalalgiaÓ. Acta Psychiat Scand 1974:46(suppl):106-113.27.Symonds CP.Aparticular variety of headache. Brain 1956;79:217-232.28.Sahota RK, Dexter JD. Sleep and headache syndromes: a clinicalreview. Headache 1990;30:80-84.2

9.Manzoni GC, Terzano MG, Bono G, et al. Cluster headacheÐclinical findings in 180 patients. Cephalalgia 1983;3:21-30.30.Russell D. Cluster headache: severity and temporal profiles ofattacks to and during 31.Dexter JD, Weitzman ED. The relationship of nocturnal headachesto sleep stage patterns. Neurology 1970;20:513-518.32.Kudrow K, McGinty DJ, Phillips ER, Stevenson M. Sleep apnea incluster headache. Cephalalgia 1984;4:33-38.33.Nappi G, Ferrari E, Polleri A, Savoldi F,Vstudy in cluster headache. Chronobiologia 1981;8:140.34.Fokin IV, Kolosova OA, Levin YI, Ossipova VV,Vein AM. Sleepcharacteristics and psychological peculiarities in cluster headachesubjects. In: Oleson J, Goadsby PJ, eds. Frontiers in HeadacheResearch 9, RelatedConditions.1999;9:196-200.35.Kudrow L. Plasma 36.Leone M, Gennaro B. Areview of hormonal findings in clusterheadache. Evidence for hypothalamic involvement. Cephalalgia1993;13:309-317.37.Chazot G, Claustrat B, Brun J, et al. Achronobiological study of LE JOURNALCANADIEN DES SCIENCES NEUROLOGIQUESVolume 29, No. 1 Ð February 200239 KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3UHVV melatonin, cortisol, growth hormone, and prolactin secretion incluster headache. Cephalalgia 1984;4:213-220.38.Waldenlind E, Gustafsson SA, Ekbom K, Wetterberg L. Circadiansecretion of cortisol and melatonin in cluster headache duringactive cluster periods and Neurol NeurosurgPsychiatr 1987;50:207-213.39.Leone M, Lucini V, DÕAmico D, et al. Twenty four hour melatoninand cortisol plasma levels in relation to timing of clusterheadache. Cephalalgia 1995;15:224-229.40. Greenspan FS. Basic and clinical endocrinology.Third Ed. EastNorwalk: Appleton and Lange, 1991.41.Leone M, Zappacosta BM, Valentini S, Colangelo AM, Bussone G.The insulin tolerance test and the ovine corticotrophin-releasinghormone in episodic cluster 42.Leone M, Maltempo C, Gritti A, Bussone G. The insulin tolerancetest and ovine corticotrophin releasing hormone test in episodiccluster headache II: comparison with low back pain patients.Cephalalgia 1994;14:357-364.43. Buijs RM, Wortel J, Van Heerikhuize JJ, et al. Anatomical andfunctional demonstration of a multisynaptic suprachiasmaticnucleus adrenal (cortex) pathway. Eur J Neurosci 1999;11:1535-1544.44.Leone M, DÕAmico D, Moschiano F, Fraschini F, Bussone G.Melatonin versus placebo in the prophylaxis of cluster headache:a double-blind pilot study with parallel groups. Cephalalgia1996;16:494-496.45.May A, Bahra A, Buchel C, Frackowiak RSJ, Goadsby PJ.Hypothalamic activation in cluster headache attacks. Lancet1998;352:275-278.46. May A, Bahra A, Buchel C, Frackowiak RSJ, Goadsby PJ. PETandMRAfindings in cluster headache and MRAin experimentalpain. Neurology 2000;55:1328-1335.47.May A, Ashburner J, Bushel C, et al. Correlation between structuraland functional changes in brain in an idiopathic headachesyndrome. Nat Med 1999;5(7):732-733.48.Human in vivotrigeminovascular activation in cluster headache. Neuropeptidechanges effects of acute attacks 1994;117:427-434.49.Pevet P, onbehaviour: possible mediation by the central GABAergic system.Neurosci Biobehav Rev 1996;20:403-412.50.Geary GG, Krause DK, Duckles SP. Melatonin directly constrictsrat cerebral arteries through modulation of potassium channels.Am J Physiol (Heart Circ Physiol) 1997;42:H1530-H1536.51.Nagtegaal JE, Smits MG, Swart ACW, Kerkhof GA, van der MeerYG. Melatonin-responsive headache in delayed sleep phasesyndrome: preliminary observations. Headache 1998;38:303-307.52.Lewy AJ. Melatonin as a marker and phase-resetter of circadianrhythms in humans. In: Olcese J, ed. Melatonin After FourDecades. New York: Kluwer Academic/Plenum Publishers. 2000.53. Weiller C, May A, Brain stem spontaneous human migraine attacks. Nat Med 1995;1:658-660.54. Bahra A, Matharu MS, Buchel C, Fracowiak RSJ, Goadsby PJ.Brain stem activation specific to migraine headache. Lancet2001;357:1016-1017. Appendix 1: The International Headache Society criteria for clusterheadache 1 are as follows:A.At least five attacks fulfilling criteria B through D.B.Severe unilateral orbital, supraorbital and/or temporal pain lasting15-180 minutes untreated.C.Headache is associated with at least one of the following signs thathave to be present on the pain-side: 1. Conjunctival injection. 2. Lacrimation. 3. Nasal congestion. 4. Rhinorrhea. 5. Forehead and facial sweating. 6. Miosis. 7. Ptosis. 8. Eye lid oedema.D.Frequency of attacks: from one every other day to eight a day.Cluster headache may be of an episodic or chronic form. Episodiccluster headache occurs in 85% of patients. In episodic cluster, attacksoccur in periods lasting from seven days to one year separated by pain-free periods lasting 14 days or more. In chronic cluster headache,attacks occur for more than a year without remission or withremissions that last less than 14 days. Chronic cluster headache isunremitting in onset in 10%, and evolves from an episodic cluster headache pattern in 5%. 2 THE CANADIAN JOURNALOF NEUROLOGICALSCIENCES40 KWWSVGRLRUJ6 3XEOLVKHGRQOLQHE\&DPEULGJH8QLYHUVLW\3U