REVIEW PAPER

The pathophysiology of brain swelling associated with subduralhemorrhage: the role of the trigeminovascular system

Waney Squier & Julie Mack & Alex Green & Tipu Aziz

Received: 22 June 2012 /Accepted: 18 July 2012# Springer-Verlag 2012

AbstractIntroduction This paper reviews the evidence in support ofthe hypothesis that the trigeminal system mediates brainswelling associated with subdural bleeding. The trigemino-vascular system has been extensively studied in migraine; itmay play an important but under-recognized role in theresponse to head trauma. Nerve fibers originating in trigem-inal ganglion cells are the primary sensors of head traumaand, through their collateral innervation of the intracranialand dural blood vessels, are capable of inciting a cascade ofvascular responses and brain swelling. The extensive tri-geminal representation in the brainstem initiates and aug-ments autonomic responses. Blood and tissue injury in thedura incite neurogenic inflammatory responses capable ofsensitizing dural nerves and potentiating the response totrauma.Discussion The trigeminal system may provide the anatomo-physiological link between small-volume, thin subduralbleeds and swelling of the underlying brain. This physiologymay help to explain the poorly understood phenomena of“second-impact syndrome,” the infant response to subdural

bleeding (the “big black brain”), as well as post-traumaticsubdural effusions. Considerable age-specific differences inthe density of dural innervation exist; age-specific responsesof this innervation may explain differences in the brain'sresponse to trauma in the young. An understanding of thispathophysiology is crucial to the development of interventionand treatment of these conditions. Antagonists to specificneuropeptides of the trigeminal system modify brain swellingafter trauma and should be further explored as potential ther-apy in brain trauma and subdural bleeding.

Keywords Big black brain . Neurogenic inflammation .

Second-impact syndrome .Subdural hemorrhage .Traumaticbrain injury . Trigeminovascular system

Introduction

Young people can suffer very acute and severe brain swell-ing associated with an overlying subdural hemorrhage(SDH). The SDH may be thin and insubstantial while thebrain swelling is significant, not respecting arterial or ve-nous territories and sparing the deep gray nuclei and theinfratentorial structures.

“Second-impact syndrome” describes this specific pat-tern of acute hemispheric swelling beneath a thin-filmSDH in adolescents. It is usually associated with a mildsecond impact occurring days or weeks after an initialimpact [12, 49].

Duhaime coined the term “big black brain” to describe asimilar phenomenon in infants and young babies [21]. Theseauthors describe the striking amount of underlying cerebralinjury associated with the subdural hematoma and noted thatage is a critical factor. In a third of cases, the swelling andhypodensity is unilateral, corresponding to and underlying athin-film SDH (Figs. 1 and 2). Older infants and toddlers more

W. Squier (*)Neuropathology, John Radcliffe Hospital,Oxford, UKe-mail: waney.squier@ndcn.ox.ac.uk

J. MackDepartment of Radiology, Penn State Hershey Medical Center,Hershey, PA, USA

A. Green : T. AzizNuffield Department of Surgical Sciences and Department ofClinical Neurology,John Radcliffe Hospital,Oxford, UK

Childs Nerv SystDOI 10.1007/s00381-012-1870-1

often develop the unilateral form of “big black brain,” whilethe bilateral pattern occurs more commonly in very younginfants with SDH [21]. The pathology of this rapid swellingand its relationship to the dural bleeding are poorly under-stood. The volume of SDH is often too small to produce masseffect, and the blood is separated from direct contact with thebrain by the arachnoid barrier layer; other indirect mecha-nisms of brain swelling need to be considered.

The distinctive pattern of swelling in both big black brainand second-impact syndrome, particularly when unilateraland sparing the deep gray structures and hindbrain, arguesagainst a central or metabolic cause. Since this anatomicalterritory corresponds to the territory innervated by the tri-geminal nerve, we hypothesize that activation of the trigem-inal system produces a vasoactive response that contributesto the severe brain swelling seen beneath thin-film SDH insome patients. Below, we present the evidence supportingthis role for the trigeminal system.

The trigeminal system

The trigeminal nerve is the largest cranial nerve and suppliessensory fibers to the face, eye, nose, mouth, skull and itsperiosteum, dura, leptomeninges, and intracranial bloodvessels. Motor branches control the muscles involved inchewing, sucking, and swallowing [78, 79]. The dura ofthe anterior and middle cranial fossae derives its sensorysupply entirely from the trigeminal nerve, while the infra-tentorial dura receives contributions from the ninth and tenthcranial nerves and the first three cervical nerves [79].

The sensory fibers also supply the blood vessels of thedural and subarachnoid compartments (Fig. 3). This vascu-lar distribution is crucial; the trigeminal provides the senso-ry supply to the cerebral blood vessels and, as such, iscapable of directly mediating vascular responses and inter-acting with both the sympathetic and parasympathetic sys-tems [29]. Each trigeminal ganglion sends fibers to theipsilateral internal carotid artery, the circle of Willis, therostral basilar artery, as well as the contralateral anteriorcerebral artery. The vessels supplying the deep gray nucleiand the caudal basilar artery receive minimal trigeminalinnervation. The midline dura and the superior sagittal sinusreceive fibers from both sides [2, 27, 63].

In addition to the innervation of the intracranial cerebralvasculature, the trigeminal fibers also innervate the menin-geal vessels and dural venous sinuses. At their entry into the

Fig. 1 A 20-month-old baby who fell from 6 ft. There is an acuteright-sided subdural hematoma. The swelling of the hemisphere be-neath the SDH is much greater than the swelling of the contralateralhemisphere. The midline shift is due to a combination of the ipsilateralswelling and mass effect from the subdural collection

Fig. 2 a and b Diffusion-weighted images in a 6-month-old whopresented with right subdural hematoma and retinal hemorrhages.Increased signal intensity is present in the territory supplied by theright trigeminal nerve, sparing the basal ganglia but involving a limitedarea of the anterior cerebral artery territory on the left side. Restricteddiffusion was confirmed on the ADC map (not shown)

Childs Nerv Syst

dural sinuses, bridging veins are equipped with innervatedsphincters which respond to alterations in intracranial pres-sure (ICP) and are involved in the control of venous bloodflow out of the brain [3, 67, 81, 84]. Thus, the trigeminalmay perceive and participate in responses to changes in boththe vessels supplying blood to the brain and those drainingblood from the brain (Fig. 4).

In addition to supplying the vasculature, the trigeminalsends fibers to the substance of the dura, arachnoid, and pia.The physiologic function of these nerve endings is not wellunderstood. Free nerve endings within the dura are thoughtto be mechanoreceptors [1] and so can be stimulated bydeformation of the skull, particularly the thin, uncalcifiedinfant skull. Chemoreceptor and mechanoreceptor nerveendings are found on the arachnoid trabeculae and the pia.These leptomeningeal sensory nerves respond to stretch aswell as changes in pH, osmolality, or inflammatory media-tors [59, 74]. The electrophysiological properties of thedural trigeminal afferents are quite different from those ofthe afferents serving the soft tissues of the head; clinically, itis recognized that the afferents serving the dura allow onlythe perception of pain [60], while stimulation of trigeminalafferents serving the soft tissues and muscles of the headcauses a wider range of sensory perception. Dural afferentsare more excitable than afferents from the temporalis muscleand more readily sensitized by inflammatory mediators [33].

Although the function of the nonvascular trigeminal inner-vation is not well understood, the anatomic arrangement issuch that sensory axons of one compartment (the integumentskull and dura) arise from the same neuronal cell bodies asaxons that are capable of releasing vasoactive peptides into aseparate compartment (the brain). Cell bodies in the trigeminal

ganglion extend processes to both the dura and the intracranialblood vessels; indeed, any single fiber innervating an intra-cranial vessel is likely to have a collateral branch supplyingthe dura [52]. This is an important anatomical construct: nervefibers in one compartment can directly affect the tissue inanother via antidromic axon–axon reflexes, even without cen-tral involvement. For example, axon–axon peripheral trigem-inal reflexes have been shown to mediate cerebrovascularadaptations to hypertension and seizures in animals after thetrigeminal ganglion has been severed from its connection tothe brain by rhizotomy [7, 63].

The effect of this direct anatomical connectivity on thebrain's response to dural bleeding has not been investigated;however, given the function of the sensory system in thedetection of, and response to, potentially harmful stimuli,the role of the trigeminal system in head trauma may beconsiderable.

The sheer size of the central representation of the trigem-inal system suggests its functional importance. It occupies asignificant part of the brainstem extending from the mid-brain through the medulla into the upper cervical spine.Collaterals from the incoming sensory fibers project to themotor nuclei of cranial nerves VII, IX, and X, forming thebasis for a number of reflexes including corneal and sneez-ing reflexes and the oxygen-conserving reflexes [30, 79].

The trigeminal system is not mature at birth and, in therat, it undergoes considerable anatomical modification andpruning throughout postnatal development [36]. Humanstudies indicate that trigeminal sensory and motor pathwayscontinue to mature throughout childhood and are instrumen-tal in the production of complex oromotor behavior, such asfeeding, swallowing, and speech [5, 26]. Our own studies of

Fig. 3 Nerves in the dura. aThere is a large nerve bundleand a smaller nerve which isclose to a venule (V). b Nerveendings are seen just beneaththe endothelium of a dural sinus(S) (stained withantineurofilament antibodies).Dural nerves stained with cCGRP and d SP

Childs Nerv Syst

human dura have shown significant variation in the densityof innervation with age. Nerve density increases between31 weeks of gestation and term, with a subsequent decreasein the first five postnatal months, remaining stable thereafter[16]. These maturational changes in the trigeminal systemmay contribute to the differences in the response of theimmature brain to traumatic injury.

In summary, the trigeminal system provides direct neuralconnection between the sensory supply of the integument ofthe head, the skull bones, the dura and leptomeninges, and thenerve supply to the blood vessels of the brain. It is thenociceptor of the brain, capable of perceiving real or impend-ing tissue damage and inciting cerebrovascular responses.This anatomical arrangement may be very important in un-derstanding the relationship between dural stimulation bytrauma or bleeding and reflex vascular responses and swellingin the cerebral hemisphere beneath it.

Trigeminal effects on cerebral vasculature

The exquisitely sensitive focal response of blood flow toneuronal activity is well known and forms the basis for func-tional brain imaging studies. This tightly regulated vascularresponse is made possible by the complex interactions be-tween neurons, glia, and cerebral blood vessels [10]. Thesympathetic and parasympathetic contributions to the controlof cerebral blood flow have been well described. The effectsof trigeminal stimulation on neural control of the circulationhave been evaluated in the context of migraine, but only rarelyexamined in other pathophysiologic conditions [29].

Stimulation of the trigeminal nerve results in the releaseof vasoactive neuropeptides including calcitonin gene-related peptide (CGRP), substance P (SP), and neurokininA. CGRP is the most potent vasodilator, while SP is respon-sible for increased vascular permeability [19] (see Fig. 3).

Fig. 4 Diagrammatic representation of the organization of the inner-vation of the meninges and their blood vessels. The dura consists oftwo leaflets, periosteal and meningeal, with tributaries of the venoussinuses between them. The meningeal arteries and veins tend to befound in the periosteal dura, but a rich capillary network is found in alllayers. The dural border layer is formed by loosely adherent cellswhich abut directly onto the arachnoid barrier membrane. The lepto-meninges (green) consist of the avascular arachnoid barrier membrane,the arachnoid trabeculae, and the pia, which covers the surface of the

brain. The arachnoid trabeculae cross the subarachnoid space andinvest the large cerebral arteries and veins within it. Capillaries arenot found in the subarachnoid space but are present in the pia togetherwith arterioles and postcapillary venules. The innervation of themeninges is represented in purple. The trigeminal ganglion sendsaxons to all layers of the meninges and the blood vessels within them.These sensory nerve fibers detect changes in the local environment andrelease neuropeptides that act on the cerebral vessels directly andindirectly

Childs Nerv Syst

The trigeminal system can produce changes in cerebral anddural blood flow by direct action on the cerebral vessels andby recruitment of parasympathetic activity via its brainstemconnections [7, 29].

Cortical spreading depression (CSD) provides a model ofthe central role of the trigeminal system in cerebrovascularresponses [10]. Activation of trigeminal afferents in the piaby CSD leads to increased blood flow and plasma extrava-sation in the overlying dura [50]. CSD occurs in traumaticand ischemic brain injury as well as subdural and corticalhemorrhage [68].

The trigeminal system is not thought to play a signif-icant role in the regulation of blood flow under restingconditions. Rather, the system acts in times of stress andhas been described as the “watchdog.” It receives earlywarning of impending threats to the circulation andresponds by mediating vasodilatation without the actualrequirement of deprivation of metabolic fuel [29]. Thiscapacity to increase cerebral blood flow independent ofmetabolic demand forms the basis for the endogenousneuroprotective “oxygen-sparing” reflexes [65].

Trauma and brain swelling

Trauma causes brain damage in two stages. In the first stage,primary mechanical deformation leads to altered membranepermeability and disturbances in ionic fluxes which, if sus-tained, lead to edema and brain swelling [42]. The secondstage is a complex cascade of events that evolves over hoursor days, when release of neurotransmitters and neuropepti-des promotes secondary tissue damage, notably throughneurogenic inflammation [19]. Intervention in this second-ary cascade is the target of current therapies.

Brain swelling is the most important determinant of mor-tality and morbidity after head trauma. Edema has tradition-ally been categorized as either vasogenic or cytotoxic [42],and in most clinical situations, there is a combination of thetwo [55]. Vasogenic edema results from the movement ofwater across the capillary walls into the perivascular spaceand then into the interstitium of the brain and leads to anincrease in brain water content, tissue swelling, and ICP. Itrequires a vascular contribution and active blood flow. Cy-totoxic edema is characterized by an increase in watercontent within cells, is fundamentally a shift of water fromthe extracellular to intracellular compartment, and by itselfdoes not result in a net increase in brain water content orbrain swelling [18].

The onset of edema may be very soon after injury; inexperimental models, blood–brain barrier (BBB) permeabil-ity is at its maximum 20 min after trauma [13]. In man,imaging studies indicate early and transient increase in BBBpermeability lasting some 60 min after trauma followed by

the development of cytotoxic edema which is greatest at 24–72 h [42–44].

Traumatic brain swelling may also result from increasedblood volume (reactive hyperemia); this phenomenon maybe particularly important in children [8, 46]. Venous con-gestion has also been proposed as the cause of unilateralswelling in traumatic brain injury (TBI) [47].

The significance of dural bleeding

In life, the dura is adherent to the underlying arachnoidbarrier membrane, the two forming a single functional unit[32]. A “subdural” hemorrhage originates in the substanceof the dura and is more precisely described as intradural;when the volume of blood is great enough, the looselyadherent cells of the dural border cell layer are cleaved,and the blood accumulates in the newly created subdural“space.” Intradural bleeding occurs at all ages, and while itis more frequent and more extensive under 6 months of age[14, 15, 70], it is an almost invariable finding when there issubdural bleeding at any age. All dural bleeding is superfi-cial to the metabolically active arachnoid barrier membraneand so separated from direct contact with the brain surface.Because the dura is anatomically distinct from the subarach-noid compartment, the trigeminal system, which providesthe neural connection between the two, becomes the primecandidate to mediate the effects of bleeding within the duraon the underlying brain.

A number of observations have shown that blood,even of small volume, in the subdural compartment cancause alterations in brain perfusion. Deibler [17] usedarterial spin labeling to show cortical hyperperfusionadjacent to subdural hematoma. Duhaime [21] showedthat blood injected into the subdural compartment in ratsled to underlying brain damage while blood layered overthe cortex did not. Specific blood constituents in subduralcollections are crucial determinants of brain swelling andstimulate trigeminal nociceptors [4, 22, 58]; injected au-tologous blood into rat dura showed alterations in bloodflow, cortical edema, and infarction of far greater magni-tude than that seen after a similar volume of blood wasinjected directly into the brain parenchyma. The largerarea of damage resulting from intradural blood, which isnot in direct contact with the brain parenchyma, indicatesthat the dura somehow modifies and enhances the paren-chymal vascular response to extravascular blood.

Damaged cells and extravasation of blood and blood prod-ucts may, of themselves, produce secondary release of pro-inflammatory neuropeptides [76]. This may explain thefamiliar inflammatory response to blood within the dura: itinduces the accumulation of macrophages and proliferation ofendothelial cells and fibroblasts which lead to the formation of

Childs Nerv Syst

the characteristic granulating membrane of chronic subduralhematoma [28, 69] (Fig. 5). CGRP released from dural nervesmay facilitate angiogenesis [77] and contribute to the prolif-eration of blood vessels seen following subdural bleeding.Heparin release from perivascular mast cells might impairhemostasis [40] and contribute to the focal rebleeding charac-teristic of the healing SDH [28].

Dural hemorrhage is associated with an increase in duralmast cell density, and the density increases as the hematomaages [80]. Mast cells appear to be important players in medi-ating the effects of dural injury on the brain. Mast cells in thebrain and meninges are early responders to cerebral ischemiaand hemorrhage. They release preformed vasoactive media-tors which promote BBB damage, brain edema, prolongedextravasation, and hemorrhage [40]. Minor traumatic damageto the skull with the dura intact is capable of activating mastcells and causing a highly localized increase in permeability ofunderlying cortical blood vessels and increased cortical hista-mine levels [73]. Simple craniotomy can cause mast celldegranulation, and mast cell activation induces a prolongedstate of excitation in trigeminal meningeal nociceptors [39].Our observation of increased mast cell density in subduralbleeding [80] provides a further clue to the mechanism bywhich SDH may activate the trigeminal system.

Bleeding into the dura also causes an alteration in ex-pression of the neuropeptides SP and CGRP in dural nerves.A significantly greater variability in response to dural bleed-ing is seen in young patients compared to adults (Davidson,unpublished results). A key to the effect of dural bleeding onthe underlying brain may lie in its ability to activate trigem-inal nociceptors and mast cells and to stimulate neurogenicinflammation.

Neurogenic inflammation

Neurogenic inflammation is a neurally elicited reactionthat produces vasodilation, increased microvascular per-meability, protein extravasation, and tissue swelling.The stimulus may be local depolarization of nerveendings or antidromic axonal reflex. Studies of periph-eral nerves show that neurogenic inflammation is theresult of stimulation of small-diameter sensory nerves(C fibers) and is mediated by the release of neuropep-tides [61]. Neurogenic inflammation has been implicat-ed in chronic systemic inflammatory diseases such asasthma and arthritis and, more specifically, in migrainewhere a considerable literature has led to a new under-standing of the trigeminovascular system. A number ofagents, including mechanical stimulation, heat, cyto-kines, and autologous blood, are capable of triggeringneurogenic inflammation [23]. The trigeminal, like allsensory nerves, can evoke neurogenic inflammation,and its role in dural inflammation and migraine hasbeen well established [23, 54].

The most extensively studied of the neuropeptides re-leased by sensory nerves are SP and CGRP. SP is the mainmediator of neurogenic inflammation. It binds to the endo-thelial NK1 receptor, leading to BBB breakdown, plasmaexudation, and edema. It can potentiate this effect by bind-ing to mast cells, leading to further release of inflammatorymediators and further vascular permeability. CGRP is apotent vasodilator and potentiates the actions of SP [76].

Neurogenic inflammation has been studied in experi-mental TBI and contributes to brain swelling; a role forSP in post-traumatic vasogenic edema has been demon-strated [19, 51]. Blocking the SP neurokinin receptorNK1 reduced post-traumatic edema, cell death, and ax-onal injury, and improved the functional outcome [18,19]. Zacest [85] showed the upregulation of SP in thecerebral microvasculature and in cortical neurons andastrocytes after human TBI. This response appears tobe mediated by peripheral sensory nerves rather than viacentral connections as no upregulation of SP was iden-tified in the medullary trigeminal tract or nucleus.

An important feature of modulators of neurogenicinflammation, such as SP, is that these agents do notact by direct synaptic transmission. Instead, they arereleased at a distance from their binding sites whichoccur diffusely in the cortex [41]. Numerous studieshave shown that perivascular sensory fibers release SPwhich is bound to glial and neuronal cells [6, 19, 45,66]. Therefore, the release of SP by trigeminal sensoryfibers in the dural and subarachnoid compartments maynot only produce direct responses in the innervatedvessels but may also act directly on wider areas of thebrain.

Fig. 5 Granulation tissue of a healing subdural membrane: freshbleeding in a healing subdural membrane. The dura is the lowest layerof poorly cellular pink fibrous tissue (D). Short arrows mark severallarge blood vessels at the junction between the dura and the morecellular healing membrane (M). Above is a fresh bleed with newcapillaries (arrows) growing between red blood cells

Childs Nerv Syst

Neurogenic inflammation and sensitization

A particular feature of neurogenic inflammation is sensiti-zation, in which chronic exposure of sensory neurons toinflammatory agents causes increased excitability, loweredthreshold to firing, and augmentation of the inflammatorysymptoms [9]. Sensitization has been considered an impor-tant factor in causing migraine [54]. Some compounds re-leased in response to tissue injury may have a minimal directeffect on neurotransmitter release, but are capable of sensi-tizing the neurogenic inflammatory response [61]. Exposureof the dura to inflammatory mediators not only directlyexcited dural nerves but also enhanced their mechanicalsensitivity so that they were strongly activated by mechan-ical stimulation that had initially caused little or no response[75]. This concept may be highly relevant in the second-impact syndrome where, following head trauma, there is adisproportionate response to a second and apparently minortrauma. The majority of cases described have had small,sometimes mixed-density, unilateral SDH [12, 49]. The factthat dural bleeding induces an increase in mast cell numbersand modifies trigeminal neuropeptide expression indicates apotential for dural bleeding to stimulate and sensitize tri-geminal sensory nerve endings and to modify and augmentthe responses of the trigeminal system, including cerebro-vascular responses. This may provide a clue not only to theetiology of second-impact syndrome but also to mechanismsof focal or unilateral brain swelling in trauma.

The trigeminal response in infants and children

Children and young people appear to show more dramaticresponses to TBI than adults. In a small proportion, geneticdisorders have been identified [71]. Sakas [64] argued a rolefor the trigeminovascular reflexes in symptoms observed inchildren following mild head injury. We agree with thishypothesis and propose that dural bleeding and inflamma-tion sensitize the dural trigeminal nerve endings, augmentthe cerebrovascular response, and contribute to a more seri-ous outcome with a cascade of catastrophic brain swellingunderlying thin-film SDH.

For decades, the “big black brain” has remained aclinical conundrum. Duhaime [21] concluded that theswelling may be a neurochemically mediated event, cau-tioned against postulating a strictly vascular occlusiveinsult or global hypoxemia as the cause, and insteadsuggested that underlying localized excitotoxicity contrib-uted to the local increased perfusion. Specifically,Duhaime points to the possible role of seizures as con-tributory to the cascade of factors leading to brain swell-ing in infant trauma. The trigeminal nerve may participatein the modulation of this cascade of events. In support of

this hypothesis is the demonstration that ablation of thetrigeminal ganglion modifies the vascular responses toseizures and hypertension in the distribution of the trigem-inal nerve [63]. Most compelling for the role of thetrigeminal in this clinical scenario is the unilateral andsupratentorial distribution of swelling, sparing the deepgray matter but involving a part of the contralateral frontalregion, the territory supplied by the trigeminal nerve. Wehypothesize that the particularly prominent vascular re-sponse to trauma in young children [21] may be in parta reflection of a combination of the more luxuriant inner-vation by the trigeminal in the young brain and theimmaturity of the trigeminal system.

Brain development involves a dynamic process ofremodeling involving both cell death and retraction of nervefibers which continues into postnatal life. This has beenshown in the corticospinal system [72]. Similarly, the tri-geminal system is disproportionately large in the fetal brainwhen it has a crucial role in early protective reflexes [37]and in the development of feeding and breathing patterns. Itthen undergoes postnatal remodeling by retraction of nervesand cell death in the trigeminal ganglion [36, 53]. Duralinnervation is very conspicuous and far denser in babies andyoung children than in adults [16] (Fig. 6). We suggest thattrigeminal responses to TBI are enhanced in young peoplebefore trigeminal maturation is complete and that immatu-rity of this system may offer a partial explanation for thedifferences between reactions of the infant and the adultbrain to trauma.

Traumatic subdural effusions

Neurogenic inflammation would provide a credible expla-nation for the development of acute fluid effusions into the

Fig. 6 Tentorium: infant, 9 months. Note the very large size of thenerve bundle between the dural leaflets (neurofilament stain)

Childs Nerv Syst

subdural compartment following trauma. This is consistentwith the early observations of McConnell [48], who pro-posed that effusions arose from the dura itself, analogous topleural, peritoneal, or joint effusions. Others have suggestedthat any pathologic condition at the dural border layer canlead to effusion and fluid accumulation, such as preexistingbleeding undergoing lysis, inflammation, and exudationfrom dural vessels or hyperacute hemorrhage [25, 83]. Thereis no clear distinction between subdural effusions and chron-ic subdural hemorrhagic collections; subdural effusions maybe xanthochromic or hemorrhagic and may evolve intofrank subdural hematomas [11, 57].

Traumatic effusions into the subdural compartment areconsidered a benign epiphenomenon of trauma and areusually seen in children under 2 years of age [11]. Theydo not develop immediately; in adults, they are recognizeddays or weeks after trauma [86]. In young children, only athird are identified within the first 24 h; effusions usuallydevelop within a week of trauma, sometimes taking up to16 days to reach maximum size [83].

Dural effusions also occur without trauma. They aredescribed after meningitis and neurosurgical proceduresand with venous thrombosis, and they may be idiopathic[11, 31, 38, 62, 82]. Typically sterile, the effusions seen withmeningitis are reactive, representing a response to the un-derlying inflammation from which they are separated by thearachnoid membrane. Some authors have hypothesized thattraumatic subdural collections and postmeningitic effusionsshare a common mechanism [82]. The typical effusionsassociated with meningitis are variable in protein contentand may be blood tinged. We propose that, in traumaticeffusions, as in postinfectious sterile collections, some ofthe fluid may come from the dura itself, either from neuro-genic vascular exudation or from intrinsic dural fluidchannels [56, 70]. Such a proteinaceous exudate into thedura could provoke additional, osmotically-driven, fluidaccumulation and/or disruption of the normal pathways ofcerebrospinal fluid (CSF) absorption through the dura, lead-ing to enlargement of the exudates.

An alternative theory for the development of traumaticeffusions is physical damage to the arachnoid barriermembrane. This hypothesis suggests that trauma mechan-ically damages the arachnoid barrier membrane, and CSFtraverses the damaged membrane collecting in the subdur-al compartment [86]. There are several reasons why thismechanism is untenable. If damage to the arachnoid bar-rier membrane existed, there would be no anatomic barrierpreventing fluid and blood from the dura from flowinginto the more capacious subarachnoid compartment. Aball-valve mechanism has been proposed to explain thepreferential movement of fluid from the subarachnoidspace into the subdural compartment through an arachnoidtear. Such a structure is hard to reconcile with the

anatomy of the thin arachnoid barrier membrane. In addi-tion, the effusions do not always colocalize with areas oftraumatic brain damage, and they occur when no commu-nication with the subarachnoid space can be demonstrated[48]. Hasegawa [34] demonstrated the passage of an in-travenous radioisotope into subdural effusions, but couldnot demonstrate passage of a radioisotope directly fromthe subarachnoid space into the effusions. Zouros [87]recovered radioactive tracer from subdural collections fol-lowing injection into the subarachnoid space. He assumedthat the tracer had passed from the subarachnoid spaceinto the dural collection through an arachnoid tear. How-ever, his patients had all undergone percutaneous subduraldrain placement which introduced a possible confoundingiatrogenic injury to the arachnoid membrane. More impor-tantly, the tracer was not found in the subdural compart-ment until after 19 h. The labeled indium he used isnormally transported across the arachnoid barrier layerinto the dura and venous sinuses, so its accumulationwithin dural collections likely represents the result of thisphysiological process. Not only is there little experimentalevidence to support the arachnoid tear hypothesis, but it ishard to imagine why fluid should pass from a largesubarachnoid space into a tissue compartment as pressuregradients would not favor this pathway.

While the etiology of post-traumatic subdural effusionsremains poorly understood, neurogenic inflammation of thedura with subsequent plasma extravasation and osmoticallydriven fluid accumulation provides an anatomically plausi-ble explanation.

Conclusion

We postulate that the trigeminovascular system has a centralrole in responding to trauma. The specific anatomy of thetrigeminal innervation, whereby single cells in the trigemi-nal ganglion send divergent processes to both the dura andthe intracranial vessels, places it in a unique position tomediate cerebral vascular responses to dural injury andbleeding. Trigeminovascular responses may be the centraland crucial mechanism linking small-volume SDH and un-derlying brain swelling.

Sensitization of the trigeminal system by prior episodesof trauma, bleeding, or inflammation may augment thesecerebrovascular responses. Sensitization combined withimmaturity of the trigeminal network may help explainphenomena such as the big black brain and second-impactsyndrome, which are typically seen in younger patients.The use of antagonists to the specific neuropeptides re-leased by trigeminal nerve stimulation or to inflammatorymediators is a therapeutic approach worthy of furtherstudy [20, 24, 35].

Childs Nerv Syst

Acknowledgments We wish to thank those families who gave usconsent to study the brains of their babies in order to undertake this work.We thankMrs. Carolyn Sloan and Joseph Davidson for staining the duralsections and Martha Hansell for drawing Fig. 4. We acknowledge ourcolleagues, particularly Professor Margaret Esiri and Dr. Monika Hofer,whose comments were most helpful in the preparation of the manuscript.

References

1. Andres KH, Von DM, Muszynski K, Schmidt RF (1987) Nervefibres and their terminals of the dura mater encephali of the rat.Anat Embryol (Berl) 175:289–301

2. Arbab MA, Wiklund L, Svendgaard NA (1986) Origin and distri-bution of cerebral vascular innervation from superior cervical,trigeminal and spinal ganglia investigated with retrograde andanterograde WGA-HRP tracing in the rat. Neuroscience 19:695–708

3. Auer LM, Ishiyama N, Hodde KC, Kleinert R, Pucher R (1987)Effect of intracranial pressure on bridging veins in rats. J Neuro-surg 67:263–268

4. Baechli H, Behzad M, Schreckenberger M, Buchholz HG, HeimannA, Kempski O, Alessandri B (2010) Blood constituents trigger brainswelling, tissue death, and reduction of glucose metabolism earlyafter acute subdural hematoma in rats. J Cereb Blood Flow Metab30:576–585

5. Barlow SM (2009) Central pattern generation involved in oral andrespiratory control for feeding in the term infant. Curr Opin Oto-laryngol Head Neck Surg 17:187–193

6. Black PH (2002) Stress and the inflammatory response: a reviewof neurogenic inflammation. Brain Behav Immun 16:622–653

7. Bolay H, Reuter U, Dunn AK, Huang Z, Boas DA, MoskowitzMA (2002) Intrinsic brain activity triggers trigeminal meningealafferents in a migraine model. Nat Med 8:136–142

8. Bruce DA, Alavi A, Bilaniuk L, Dolinskas C, Obrist W, Uzzell B(1981) Diffuse cerebral swelling following head injuries in chil-dren: the syndrome of “malignant brain edema”. J Neurosurg54:170–178

9. Burstein R, Jakubowski M (2004) Analgesic triptan action in ananimal model of intracranial pain: a race against the developmentof central sensitization. Ann Neurol 55:27–36

10. Busija DW, Bari F, Domoki F, Horiguchi T, Shimizu K (2008)Mechanisms involved in the cerebrovascular dilator effects ofcortical spreading depression. Prog Neurobiol 86:379–395

11. Caldarelli M, Di RC, Romani R (2002) Surgical treatment ofchronic subdural hygromas in infants and children. Acta Neurochir(Wien) 144:581–588

12. Cantu RC, Gean AD (2010) Second-impact syndrome and a smallsubdural hematoma: an uncommon catastrophic result of repetitivehead injury with a characteristic imaging appearance. J Neuro-trauma 27:1557–1564

13. Cernak I, Chang T, Ahmed FA, Cruz MI, Vink R, Stoica B, FadenAI (2010) Pathophysiological response to experimental diffusebrain trauma differs as a function of developmental age. DevNeurosci 32:442–453

14. Cohen MC, Scheimberg I (2008) Evidence of occurrence of intra-dural and subdural hemorrhage in the perinatal and neonatal periodin the context of hypoxic ischemic encephalopathy: an observa-tional study from two referral institutions in the United Kingdom.Pediatr Dev Pathol 12:169–176

15. Cohen MC, Sprigg A, Whitby EH (2010) Subdural hemorrhage,intradural hemorrhage and hypoxia in the pediatric and perinatalpost mortem: are they related? An observational study combiningthe use of post mortem pathology and magnetic resonance imag-ing. Forensic Sci Int 200:100–107

16. Davidson JR, Mack J, Gutnikova A, Varatharaj A, Darby S, SquierW (2012) Developmental changes in human dural innervation.Childs Nerv Syst 28:665–671

17. Deibler AR, Pollock JM, Kraft RA, Tan H, Burdette JH, MaldjianJA (2008) Arterial spin-labeling in routine clinical practice, part 3:hyperperfusion patterns. AJNR AmJ Neuroradiol 29:1428–1435

18. Donkin JJ, Vink R (2010) Mechanisms of cerebral edema intraumatic brain injury: therapeutic developments. Curr Opin Neu-rol 23:293–299

19. Donkin JJ, Nimmo AJ, Cernak I, Blumbergs PC, Vink R (2009)Substance P is associated with the development of brain edema andfunctional deficits after traumatic brain injury. J Cereb Blood FlowMetab 29:1388–1398

20. Donkin JJ, Cernak I, Blumbergs PC, Vink R (2011) A substance Pantagonist reduces axonal injury and improves neurologic outcomewhen administered up to 12 h after traumatic brain injury. J Neuro-trauma 28:217–224

21. Duhaime AC, Durham S (2007) Traumatic brain injury in infants:the phenomenon of subdural hemorrhage with hemispheric hypo-density (“big black brain”). Prog Brain Res 161:293–302

22. Ebersberger A, Handwerker HO, Reeh PW (1999) Nociceptiveneurons in the rat caudal trigeminal nucleus respond to bloodplasma perfusion of the subarachnoid space: the involvement ofcomplement. Pain 81:283–288

23. Eikermann-Haerter K, Moskowitz MA (2008) Animal models ofmigraine headache and aura. Curr Opin Neurol 21:294–300

24. Elliott MB, Tuma RF, Amenta PS, Barbe MF, Jallo J (2011) Acuteeffects of a selective cannabinoid-2 receptor agonist on neuroin-flammation in a murine model of traumatic brain injury. J Neuro-trauma 28:973–981

25. Feng JF, Jiang JY, Bao YH, Liang YM, Pan YH (2008) Traumaticsubdural effusion evolves into chronic subdural hematoma: twostages of the same inflammatory reaction? Med Hypotheses70:1147–1149

26. Finan DS, Smith A (2005) Jaw stretch reflexes in children. ExpBrain Res 164:58–66

27. Fricke B, Andres KH, Von DM (2001) Nerve fibers innervating thecranial and spinal meninges: morphology of nerve fiber terminalsand their structural integration. Microsc Res Tech 53:96–105

28. Friede RL (1989) Subdural haematomas, hygromas and effusions.In: Friede RL (ed) Developmental neuropathology. Springer, Got-tingen, pp 198–208

29. Goadsby PJ, Sercombe R (1996) Neurogenic regulation of cerebralblood flow: extrinsic neural control. In: Mraovitch S, Sercombe R(eds) Neurophysiological basis of cerebral blood flow control: anintroduction. John Libbey and Co Ltd, London, pp 285–321

30. Gorini C, Philbin K, Bateman R, Mendelowitz D (2010) Endoge-nous inhibition of the trigeminally evoked neurotransmission tocardiac vagal neurons by muscarinic acetylcholine receptors. JNeurophysiol 104:1841–1848

31. Guthkelch AN (1953) Subdural effusions in infancy: 24 cases. BrMed J 1:233–239

32. Haines DE, Harkey HL, Al Mefty O (1993) The “subdural” space:a new look at an outdated concept. Neurosurgery 32:111–120

33. Harriott AM, Gold MS (2009) Electrophysiological properties ofdural afferents in the absence and presence of inflammatory medi-ators. J Neurophysiol 101:3126–3134

34. Hasegawa M, Yamashima T, Yamashita J, Suzuki M, Shimada S(1992) Traumatic subdural hygroma: pathology and meningealenhancement on magnetic resonance imaging. Neurosurgery31:580–585

35. Ho TW, Ferrari MD, Dodick DW, Galet V, Kost J, Fan X,Leibensperger H, Froman S, Assaid C, Lines C, Koppen H,Winner PK (2008) Efficacy and tolerability of MK-0974(telcagepant), a new oral antagonist of calcitonin gene-relatedpeptide receptor, compared with zolmitriptan for acute migraine:

Childs Nerv Syst

a randomised, placebo-controlled, parallel-treatment trial. Lancet372:2115–2123

36. Horgan K, O’Connor TP, van der Kooy D (1990) Prenatal speci-fication and target induction underlie the enrichment of calcitoningene-related peptide in the trigeminal ganglion neurons projectingto the cerebral vasculature. J Neurosci 10:2485–2492

37. Humphrey T (1952) The spinal tract of the trigeminal nerve inhuman embryos between 7 1/2 and 8 1/2 weeks of menstrualage and its relation to early fetal behavior. J Comp Neurol97:143–209

38. Krasnokutsky MV (2011) Cerebral venous thrombosis: a potentialmimic of primary traumatic brain injury in infants. Am J Roent-genol 197:W503–W507

39. Levy D, Burstein R, Kainz V, Jakubowski M, Strassman AM(2007) Mast cell degranulation activates a pain pathway underly-ing migraine headache. Pain 130:166–176

40. Lindsberg PJ, Strbian D, Karjalainen-Lindsberg ML (2010) Mastcells as early responders in the regulation of acute blood–brainbarrier changes after cerebral ischemia and hemorrhage. J CerebBlood Flow Metab 30:689–702

41. Liu H, Brown JL, Jasmin L, Maggio JE, Vigna SR, Mantyh PW,Basbaum AI (1994) Synaptic relationship between substance P andthe substance P receptor: light and electron microscopic character-ization of the mismatch between neuropeptides and their receptors.Proc Natl Acad Sci USA 91:1009–1013

42. Marmarou A (2007) A review of progress in understanding thepathophysiology and treatment of brain edema. Neurosurg Focus22:E1

43. Marmarou A, Signoretti S, Aygok G, Fatouros P, Portella G (2006)Traumatic brain edema in diffuse and focal injury: cellular orvasogenic? Acta Neurochir 96:24–29

44. Marmarou A, Signoretti S, Fatouros PP, Portella G, Aygok GA,Bullock MR (2006) Predominance of cellular edema in traumaticbrain swelling in patients with severe head injuries. J Neurosurg104:720–730

45. Marriott I, Bost KL (1998) Substance P diminishes lipopolysac-charide and interferon-gamma-induced TGF-beta 1 production bycultured murine macrophages. Cell Immunol 183:113–120

46. Martin NA, Patwardhan RV, Alexander MJ, Africk CZ, Lee JH,Shalmon E, Hovda DA, Becker DP (1997) Characterization ofcerebral hemodynamic phases following severe head trauma:hypoperfusion, hyperemia, and vasospasm. J Neurosurg 87:9–19

47. Matsushige T, Kiya K, Satoh H, Mizoue T, Kagawa K, Araki H(2004) Cerebral venous congestion following closed head injury ina child. Pediatr Neurosurg 40:241–244

48. McConnell AA (1941) Traumatic subdural effusions. J NeurolPsychiatr 4:237–256

49. Mori T, Katayama Y, Kawamata T (2006) Acute hemisphericswelling associated with thin subdural hematomas: pathophysiol-ogy of repetitive head injury in sports. Acta Neurochir 96:40–43

50. Moskowitz MA (2007) Genes, proteases, cortical spreading de-pression and migraine: impact on pathophysiology and treatment.Funct Neurol 22:133–136

51. Nimmo AJ, Cernak I, Heath DL, Hu X, Bennett CJ, Vink R (2004)Neurogenic inflammation is associated with development of ede-ma and functional deficits following traumatic brain injury in rats.Neuropeptides 38:40–47

52. O’Connor TP, van der Kooy D (1986) Pattern of intracranial andextracranial projections of trigeminal ganglion cells. J Neurosci6:2200–2207

53. O’Connor TP, van der Kooy D (1986) Cell death organizes thepostnatal development of the trigeminal innervation of the cerebralvasculature. Brain Res 392:223–233

54. Olesen J, Burstein R, Ashina M, Tfelt-Hansen P (2009) Origin ofpain in migraine: evidence for peripheral sensitisation. LancetNeurol 8:679–690

55. Papadopoulos MC, Verkman AS (2007) Aquaporin-4 and brainedema. Pediatr Nephrol 22:778–784

56. Papaiconomou C, Zakharov A, Azizi N, Djenic J, Johnston M(2004) Reassessment of the pathways responsible for cerebrospinalfluid absorption in the neonate. Childs Nerv Syst 20:29–36

57. Park CK, Choi KH, Kim MC, Kang JK, Choi CR (1994) Sponta-neous evolution of posttraumatic subdural hygroma into chronicsubdural haematoma. Acta Neurochir (Wien) 127:41–47

58. Patel TR, Schielke GP, Hoff JT, Keep RF, Lorris BA (1999)Comparison of cerebral blood flow and injury following intrace-rebral and subdural hematoma in the rat. Brain Res 829:125–133

59. Raja SN, Meyer RA, Ringkamp M, Campbell JN (1999) Periph-eral neural mechanisms of nociception. In: Wall PD, Melzack R(eds) Textbook of pain. Churchill Livingstone, Edinburgh

60. Ray BWH (1940) Experimental studies on headache. Arch Surg41:813–856

61. Richardson JD, Vasko MR (2002) Cellular mechanisms of neuro-genic inflammation. J Pharmacol Exp Ther 302:839–845

62. Roach ES, Golomb MR, Adams R, Biller J, Daniels S, DeVeber G,Ferriero D, Jones BV, Kirkham FJ, Scott RM, Smith ER (2008)Management of stroke in infants and children: a scientific state-ment from a Special Writing Group of the American Heart Asso-ciation Stroke Council and the Council on Cardiovascular Diseasein the Young. Stroke 39:2644–2691

63. Sakas DE, Moskowitz MA, Wei EP, Kontos HA, Kano M,Ogilvy CS (1989) Trigeminovascular fibers increase blood flowin cortical gray matter by axon reflex-like mechanisms duringacute severe hypertension or seizures. Proc Natl Acad Sci USA86:1401–1405

64. Sakas DE, Whittaker KW, Whitwell HL, Singounas EG (1997)Syndromes of posttraumatic neurological deterioration in childrenwith no focal lesions revealed by cerebral imaging: evidence for atrigeminovascular pathophysiology. Neurosurgery 41:661–667

65. Sandu N, Spiriev T, Lemaitre F, Filis A, Schaller B (2010) Newmolecular knowledge towards the trigemino-cardiac reflex as acerebral oxygen-conserving reflex. Sci World J 10:811–817

66. Severini C, Improta G, Falconieri-Erspamer G, Salvadori S,Erspamer V (2002) The tachykinin peptide family. PharmacolRev 54:285–322

67. Si Z, Luan L, Kong D, Zhao G, Wang H, Zhang K, Yu T, Pang Q(2008) MRI-based investigation on outflow segment of cerebralvenous system under increased ICP condition. Eur J Med Res13:121–126

68. Smith JM, Bradley DP, James MF, Huang CL (2006) Physiologicalstudies of cortical spreading depression. Biol Rev Camb PhilosSoc 81:457–481

69. Squier W, Mack J (2009) The neuropathology of infant subduralhaemorrhage. Forensic Sci Int 187:6–13

70. Squier W, Lindberg E, Mack J, Darby S (2009) Demonstration offluid channels in human dura and their relationship to age andintradural bleeding. Childs Nerv Syst 25:925–931

71. Stam AH, Luijckx GJ, Poll-The BT, Ginjaar IB, Frants RR, Haan J,Ferrari MD, Terwindt GM, van den Maagdenberg AM (2009)Early seizures and cerebral oedema after trivial head trauma asso-ciated with the CACNA1A S218L mutation. J Neurol NeurosurgPsychiatry 80:1125–1129

72. Staudt M (2010) Brain plasticity following early life brain injury:insights from neuroimaging. Semin Perinatol 34:87–92

73. Stokely ME, Orr EL (2008) Acute effects of calvarial damage ondural mast cells, pial vascular permeability, and cerebral corticalhistamine levels in rats and mice. J Neurotrauma 25:52–61

74. Strassman AM, Levy D (2006) Response properties of dural noci-ceptors in relation to headache. J Neurophysiol 95:1298–1306

75. Strassman AM, Raymond SA, Burstein R (1996) Sensitization ofmeningeal sensory neurons and the origin of headaches. Nature384:560–564

Childs Nerv Syst

76. Thornton E, Ziebell JM, Leonard AV, Vink R (2010) Kinin recep-tor antagonists as potential neuroprotective agents in central ner-vous system injury. Molecules 15:6598–6618

77. Toda M, Suzuki T, Hosono K, Kurihara Y, Kurihara H, Hayashi I,Kitasato H, Hoka S, Majima M (2008) Roles of calcitonin gene-related peptide in facilitation of wound healing and angiogenesis.Biomed Pharmacother 62:352–359

78. Troger J, Kieselbach G, Teuchner B, Kralinger M, Nguyen QA,Haas G, Yayan J, Gottinger W, Schmid E (2007) Peptidergicnerves in the eye, their source and potential pathophysiologicalrelevance. Brain Res Rev 53:39–62

79. Truex RC, Carpenter MB (1969) Human neuroanatomy. The Wil-liams and Wilkins Company, Baltimore

80. Varatharaj A, Mack J, Davidson JR, Gutnikov A, Squier W (2012)Mast cells in the human dura: effects of age and dural bleeding.Childs Nerv Syst 28:541–545

81. Vignes JR, Dagain A, Guerin J, Liguoro D (2007) A hypothesis ofcerebral venous system regulation based on a study of the junctionbetween the cortical bridging veins and the superior sagittal sinus.Laboratory investigation. J Neurosurg 107:1205–1210

82. Vinchon M, Joriot S, Jissendi-Tchofo P, Dhellemmes P(2006) Postmeningitis subdural fluid collection in infants:changing pattern and indications for surgery. J Neurosurg104:383–387

83. Wells RG, Sty JR (2003) Traumatic low attenuation subdural fluidcollections in children younger than 3 years. Arch Pediatr AdolescMed 157:1005–1010

84. Yu Y, Chen J, Si Z, Zhao G, Xu S, Wang G, Ding F, Luan L,Wu L, Pang Q (2010) The hemodynamic response of thecerebral bridging veins to changes in ICP. Neurocrit Care12:117–123

85. Zacest AC, Vink R, Manavis J, Sarvestani GT, Blumbergs PC(2010) Substance P immunoreactivity increases following humantraumatic brain injury. Acta Neurochir 106:211–216

86. Zanini MA, de Lima Resende LA, de Souza Faleiros AT, GabarraRC (2008) Traumatic subdural hygromas: proposed pathogenesisbased classification. J Trauma 64:705–713

87. Zouros A, Bhargava R, Hoskinson M, Aronyk KE (2004) Furthercharacterization of traumatic subdural collections of infancy. Re-port of five cases. J Neurosurg 100:512–518

Childs Nerv Syst