Intracranial Pressure Changes following TraumaticBrain Injury in Rats: Lack of Significant Change

in the Absence of Mass Lesions or Hypoxia

Levon Gabrielian,1 Luke W. Willshire,1 Stephen C. Helps,1 Corinna van den Heuvel,1

Jane Mathias,2 and Robert Vink1

Abstract

Traumatic brain injury (TBI) often causes raised intracranial pressure (ICP), with > 50% of all TBI- related deathsbeing associated with this increase in ICP. To date, there is no effective pharmacological treatment for TBI, partlybecause widely used animal models of TBI may not replicate many of the pathophysiological responses ob-served in humans, and particularly the ICP response. Generally, rodents are the animal of choice in neurotraumaresearch, and edema formation has been demonstrated in rat models; however, few studies in rats have spe-cifically explored the effects of TBI on ICP. The aim of the current study was to investigate the ICP response ofrats in two different, focal and diffuse, injury models of TBI. Adult male Sprague-Dawley rats were subjected tobrain trauma by either lateral fluid percussion or impact-acceleration induced injury, in the presence or absenceof secondary hypoxia. ICP, mean arterial blood pressure (MABP), and cerebral perfusion pressure (CPP) weremonitored for 4 h after TBI. TBI alone or coupled with hypoxia did not result in any significant increase of ICP inrats unless there was an intracranial hemorrhage. At all other times, changes in CPP were the result of changes inMABP and not ICP. Our results suggest that rats may be able to compensate for the intracranial expansionassociated with cerebral edema after TBI, and that they only develop a consistent post-traumatic increase in ICPin the presence of a mass lesion. Therefore, they are an inappropriate model for the investigation of ICP changesafter TBI, and for the development of therapies targeting ICP.

Key words: brain swelling; fluid percussion; hemorrhage; impact-acceleration; neurotrauma

Introduction

Traumatic brain injury (TBI) is the leading cause ofdeath and disability in people < 45 years of age in Wes-

tern industrialized countries (Bruns et al., 2003; Finfer andCohen, 2001; Fleminger and Ponsford, 2005; Harris et al.,2003), with the main causes of TBI including motor vehicleaccidents, falls, assaults, and sporting activities (Langloiset al., 2006). In addition to the great social burden placed onthe families of TBI survivors, the economic burden of TBI,including gross medical costs and loss of productivity, is es-timated to be $60 billion annually in the United States alone(Langlois et al., 2006).

The pathophysiological changes that follow TBI are acomplex network of structural, physiological, and functionalsecondary changes that are initiated by the primary me-chanical event. Whereas a number of injury factors have beenindentified in this secondary cascade, brain edema develop-ment, blood–brain barrier (BBB) breakdown, and decreased

brain tissue perfusion, in particular, have been identified ascritical to outcome (Bareyre et al., 1997; Unterberg et al., 2004;Vink and Nimmo, 2009). Brain tissue edema is a critical factor,as it causes an expansion of brain volume and thus increasesintracranial pressure, consequently reducing cerebral perfu-sion and oxygenation (Marmarou 2007; McIntosh et al., 1990).

Traumatic brain edema is basically classified as either va-sogenic or cytotoxic (Donkin and Vink, 2010; Klatzo 1967).Vasogenic edema is the result of the BBB disruption and ac-cumulation of extracellular fluid, rich in protein, thus in-creasing the volume of the interstitial space without cellswelling (Gardenfors et al., 2002; Papadopoulos et al., 2004).In contrast, cytotoxic edema results from the sustained in-tracellular water accumulation, rich in electrolytes, mainlycaused by active ion pump failure (e.g., Na + /K + -ATPase),resulting in cell swelling and decreased interstitial space(Liang et al., 2007). In reality, the edema that occurs after TBIis more often a mix of the two types with several sub-types having been suggested to differentiate between the

Schools of 1Medical Sciences and 2Psychology, Adelaide Centre for Neurological Diseases, University of Adelaide, Adelaide SA, Australia.

JOURNAL OF NEUROTRAUMA 28:2103–2111 (October 2011)ª Mary Ann Liebert, Inc.DOI: 10.1089/neu.2011.1785

2103

pathogenic origins (Katayama and Kawamata, 2003; Unter-berg et al., 2004). In any case, various mediators, such asglutamate, lactate, nitric oxide, arachidonic acid, substance P,and kinins are being released, which also contribute to thefurther development of both vasogenic and cytotoxic edema(Schilling and Wahl, 1999; Schneider et al., 1992), thus per-petuating the deleterious effects of brain water accumulation.

More than half of all deaths following TBI result from un-resolved secondary brain edema that causes increased intra-cranial pressure (ICP) and reduced cerebral perfusionpressure (Feickert et al 1999; Rowland and Morris, 2007; Stahlet al., 2001). This makes management of raised ICP the singlemost important treatment strategy in TBI today (Miller et al.,1977; Marmarou, 1992; Wahlstrom et al., 2005; Steiner andAndrews, 2006). Although ICP has been the focus of intensivecare management in TBI survivors for a number of decades,the mechanisms associated with brain water accumulation arelargely unknown (Marmarou et al., 1991; Soustiel and Lar-isch, 2010) and, consequently, no effective therapeutic inter-vention currently exists for secondary cerebral edema andraised ICP (Beauchamp et al., 2008; Tolias and Bullock, 2004;Vink and Nimmo, 2009).

The development of novel pharmacological treatments re-quires extensive animal research. Whereas a number of ef-fective animal models have been developed that recreatemany of the pathophysiological responses observed in hu-mans, the cost-effective and highly reproducible nature of ratmodels has made them the model of choice in neurotraumaresearch (Cernak, 2005; Lighthall et al., 1989). Although a fewstudies have looked at ICP changes after TBI in rats, no studyhad adequately investigated the effects of TBI on ICP in dif-ferent models of TBI in rats. A thorough characterization ofthis effect is of paramount importance to the future develop-ment of ICP treatment, and is the primary focus of this study.In addition, we will investigate the effects of superimposedsecondary hypoxia on ICP. In clinical scenarios, TBI is fre-quently accompanied by a reduction in the oxygen concen-tration of arterial blood (Bullock et al., 2000). This hypoxemiaresults from the apnea that accompanies the traumatic event,and is caused by failure or dysfunction of the respiratory re-sponse (Atkinson et al., 1998; Kemp et al., 2003). Hypoxemialeads to hypoxic or ischemic brain injury, which is known tosignificantly contribute to morbidity and mortality followingTBI (Stocchetti et al., 1996).

Methods

Subjects

A total of 48 adult male Sprague-Dawley rats weighing400–500 g were used in this study. All animals were housed ina conventional rodent facility at 24�C on a 12-h light/darkcycle with access to standard rodent food and water ad libitum.All experimental protocols were conducted according to theguidelines established by the Australian National Health andMedical Research Council and were approved by the animalethics committees of the Institute of Medical and VeterinaryScience and the University of Adelaide.

Experimental design

All animals were randomly assigned to the following sixexperimental groups: sham group for both models (uninjured,

no hypoxia, n = 9); lateral fluid percussion (LFP) injury only(n = 8); LFP injury with secondary hypoxia (n = 9); weight drop(WD) injury only (n = 7); WD injury with secondary hypoxia(n = 7); and hypoxia only for both models, as a control for thehypoxic episode (n = 8). Sham group animals underwent allpre-trauma surgical preparation for both models, but werenot subject to TBI. In two of the injury groups, secondaryhypoxia (10% oxygen; 90% nitrogen) was induced for 15 minimmediately following either LFP or WD in order to replicatethe secondary hypoxia often experienced in clinical scenariosthrough transient apnea. The parameters for hypoxic ventila-tion were adapted during preliminary experiments from themethods of Rolett and colleagues (Rolett et al., 2000) to achievea blood PO2 of * 45–50 mmHg. Following the 15 min hypoxicepisode, animals were ventilated with a normoxic mixture(30% O2, 70% N2) for the remainder of the monitoring period.

At the conclusion of the monitoring period, animals werekilled by decapitation under pentobarbital anesthesia andtheir brains subjected to gross neuropathological examinationto detect mass lesions. To understand the impact of bleedingon the variables under investigation, data from animals withsignificant intracranial hemorrhage were pooled and subjectto separate analysis. A subgroup of animals without brainhemorrhage (6 shams, 6 trauma, and 6 trauma plus hypoxia)were used for determination of edema.

LFP injury

LFP was induced as described in detail elsewhere(McIntosh et al., 1989, Rogatsky et al., 1996, Thompson et al.,2005). Briefly, anesthesia was induced with 4% isoflurane ina normoxic mixture of 70% N2O and 30% O2 for 3–5 min.Anesthesia was then maintained at 2% of isoflurane via a nosecone while a tracheostomy was performed and the animalssubsequently ventilated at * 225 mL/min. The animal’s rightfemoral artery and vein were cannulated using a polyethylenecannula (PE 50; inside diameter = 0.58 mm, outside diame-ter = 0.97 mm), with the arterial cannula used for monitoring ofblood pressure and sampling of arterial blood for gas analysisand the venous cannula used for urethane administration afterswitching from isoflurane to urethane for maintenance anes-thesia. Unlike isoflurane, urethane produces a long-lastingsteady level of surgical anesthesia with minimal effects onthe autonomic, respiratory, and cardiovascular systems (Fish,1997; Maggi and Meli, 1986). The animal’s core temperaturewas monitored and maintained at 37.5 – 0.5�C using a ther-mostatically heating pad. A sagittal incision (2 cm) was madeon the dorsal scalp of the head and the temporal muscles werereflected before a 5 mm in diameter craniectomy was trephinedinto the skull centered 3 mm right of the sagittal suture andmidway between the bregma and lambda. The dura was keptintact at the opening and a specially designed steel-bolt (4 mminside diameter and 5 mm outside) was screwed and securedinto the craniectomy with cyanoacrylate adhesive.

Immediately prior to trauma, animals were placed in aprone position onto a foam block and the steel bolt was filledwith isotonic saline. Then the animal was attached to the LFPdevice (McIntosh et al., 1989) and a moderate trauma wasinduced with a force of 2.6–2.8 atm. After injury, an ICP probe(Codman ICP Express) was inserted to a depth of 6 mm belowthe dura through the steel bolt and into the cerebral paren-chyma, and then sealed with sterile bone wax (Ethicon,

2104 GABRIELIAN ET AL.

W810). Arterial blood gases were analyzed periodically andarterial blood pressure was monitored continuously.

WD injury

Animals were injured using the Marmarou impact-acceleration model of diffuse TBI, which has been describedextensively elsewhere (Marmarou et al., 1994). Briefly, ratswere anesthetized, intubated, ventilated, and their core tem-perature was maintained at 37.5 – 0.5�C, as described for theLFP animals. The skull was exposed after midline sagittalincision (2 cm) and soft tissue retraction. A drop of cyanoac-rylate adhesive was then applied to the skull along the mid-line, midway between the lambda and the bregma sutures,and a protective stainless steel disc (10 mm diameter x 3 mmthick) was fixed centrally upon the skull. The stainless steeldisc acted as a helmet during trauma to reduce the incidenceof skull fractures. Immediately prior to trauma, animals weresecured in a prone position onto a foam block of uniformdensity (depth 11.5 cm) and placed under the injury appara-tus, ensuring that the brass weight was centrally alignedabove the protective steel disc. A TBI was induced by re-leasing the brass weight (450 g) from a height of 2 m via acylindrical PVC conduit and allowing it to impact onto thesteel disc attached to the animal’s skull. The animal wasrapidly relocated after the initial impact to prevent any furthercontact from the rebounding weight. The stainless steel discwas removed and the animal placed in a stereotactic frame(David Kopf Instruments, Tujunga, CA). For insertion ofthe ICP probe, a burr hole (2 mm diameter) was made in theanimal’s left parietal bone at the stereotactic coordinate,0.5 mm posterior to the bregma and 4 mm lateral to the mid-line suture. After puncturing the dura mater, the ICP probewas directly inserted 6 mm below the dura into the cerebralparenchyma.

Edema measurement

Brain water content was assessed by the wet weight–dryweight method (O’Connor et al., 2006; Donkin et al., 2009) at5 h after focal or diffuse trauma. Rats (n = 6 per group) weredecapitated under pentobarbital anesthesia and their brainsremoved. The cerebellum and brainstem were then rapidlyremoved and the cerebral hemispheres placed in glass vials.After weighing, brains were dried at 100�C for 24 h. The dryweight was then obtained and brain water content calculatedaccording to the formula

water¼ðwet weight�dry weightÞ=wet weight · 100

Blood gas analysis

Arterial blood samples (0.2 mL) were obtained via thefemoral cannula and analyzed using an Osmetech OPTI cas-sette system with an Osmetech OPTI blood gas analyzer(CCA, Helena Laboratories Pty Ltd., Australia). After TBI,blood gas analysis was conducted twice in each animal, onceat 10 min and again at 210 min post-trauma.

ICP monitoring

ICP was measured with a CODMAN� microsensorprobe, which consisted of a nylon catheter with a silicon

strain-gauge type sensor mounted at the tip (1.2 mm diam-eter; Codman & Shurtleff, Inc., DePuySpine�). The ICPprobe was connected to a Codman Express monitor(DePuySpine�) and measurements of ICP were recordedcontinuously by a MacLab data acquisition system (MacLab2e) following insertion of the probe. Probe insertion wascomplete by 60 min post-trauma.

Blood pressure monitoring

Mean arterial blood pressure (MABP) was monitored using aMacLab data acquisition system (MacLab 2e). A Statham-typepressure transducer was connected via a polyethylene tube tothe animal’s arterial cannula. The pressure trace was relayedfrom the transducer to the MacLab via a bridge amp and all datawere recorded and stored using a personal computer.

Determination of cerebral perfusion pressure

Cerebral perfusion pressure (CPP) at individual time pointswas calculated by the equation:

CPP¼MABP� ICP

and stored using a MacLab data acquisition system (MacLab2e) linked to a personal computer.

Statistical analysis

All data are shown as means and SEMs and were analyzedby two-way, repeated measures analysis of variance (ANO-VA), followed by individual Student Newman Keuls tests.A p value of < 0.05 was considered significant.

Results

Mortality and hemorrhage

The overall mortality rate for this investigation was 8% (2FPI; 2 WD), which is consistent with of < 10% mortality pre-vious reports at this severity of injury (McIntosh et al., 1989;Marmarou et al., 1994). Deaths were caused by skull fractureor respiratory failure. Mass hemorrhagic lesions were foundin WD animals that were excluded on the basis of skull frac-ture but not in animals that did not sustain a skull fracture.Following FPI, 5 animals were found to have mass hemor-rhagic lesions and were analyzed separately.

Edema

Brain water content in sham (uninjured animals) was78.5 – 0.3%. At 5 h after TBI, brain water content had increasedto 79.8 – 0.1%, whereas in injured animals exposed to a hyp-oxic episode, it had increased to 81.4 – 0.4%. These results areconsistent with previous reports that have shown that theserodent TBI models typically cause between 1 and 4% edemaformation (Donkin et al., 2009; Nida et al., 1995; O’Connoret al., 2006; Soares et al., 1992), which may be exacerbated byhypoxia (Ishige et al., 1987; Van Putten et al., 2005).

Blood gas analysis

Blood oxygen saturation (SO2) was significantly diminishedin animals subjected to TBI with hypoxia or hypoxia only(Table 1) confirming the effectiveness of the hypoxic period.

ICP CHANGES IN RATS AFTER TBI 2105

The observed decrease in SO2 was physiologically as well asstatistically significant, as the initial reading for each group fellbelow the clinical reference limit of 90%. By 210 min followingtrauma, SO2 had returned to sham levels in all groups.

Arterial blood pressure

MABP in sham and injured animals, with and withouthypoxia, is shown in Figure 1. Prior to injury, the MABPacross all animals was 94 – 3 mm Hg. During the hypoxicepisode, MABP decreased to < 60 mm Hg in both injured andnon-injured animals, and returned to near sham levels soonafter the resumption of normoxia. The return to near normalMABP after hypoxia was more gradual in the WD injuredanimals than in the LFP injured animals. After TBI, mild hy-potension was apparent over the post-traumatic monitoringperiod. Such hypotension is typical of TBI, albeit that it wasless apparent after a hypoxic episode.

ICP analysis

Sham group animals in both models of TBI demonstratedan average ICP of * 4–5 mm Hg during the 4 h post-injury. In

the groups receiving TBI only, there was no significant in-crease in ICP compared to sham groups in either FPI or WDinjury (Fig. 2), with ICP typically changing by < 2 mm Hg.Similarly, animals exposed to hypoxia (with no TBI) showed aslight increase of 1–2 mm Hg in ICP compared to sham groupanimals, which was also insignificant. However, when TBIwas combined with hypoxia, there was a slight increase in ICPin both models. In the LFP model, this increase never ex-ceeded a mean of 8 mm Hg, which was insignificant. How-ever, when secondary hypoxia was combined with WDinjury, there was a small but significant rise in ICP ( p < 0.05).ICP gradually increased over time from a mean of 8 mm Hg toa mean of 11 mm Hg at 4 h after injury. This increase in ICP,although statistically significant compared to shams, was only2–4 mm Hg above that of transient hypoxia alone ( p > 0.05),which is unlikely to be particularly deleterious.

No previous study has explored the importance of brainhemorrhage on ICP changes after TBI in rats. We had retro-spectively separated animals that had intracranial bleedingand excluded them from the above analysis. When poolingthese animals with an intracranial hemorrhage and analyzingtheir ICP changes, it was noted that ICP was significantly

Table 1. Blood Gas Analysis at 10 Min and 210 Min for Animals in the Groups Receiving TBI Alone,TBI with Hypoxia, Hypoxia Alone, or No Injury (Sham)

Sham TBI TBI with hypoxia Hypoxia

PO2 10 min 149.8 – 8.4 182.8 – 5.5* 49.2 – 1.0*** 58.3 – 4.9***210 min 112.2 – 17.4{ 170 – 5.5*** 104.8 – 4.9{{{ 131 – 13.2{{{

SO2 10 min 99.17 – 0.2 99.2 – 0.2 78.4 – 1.2*** 85.5 – 2.8***210 min 97 – 1.0 99.2 – 0.2 96.25 – 0.6{{{ 98 – 0.4{{{

PCO2 10 min 34.33 – 1.7 34.25 – 3.0 32.8 – 0.9 36.66 – 1.4210 min 32 – 2.8 24.4 – 2.2*{ 35.75 – 0.8 37.16 – 2.5

pH 10 min 7.5 – 0.02 7.5 – 0.41 7.45 – 0.01 7.49 – 0.01210 min 7.46 – 0.01 7.51 – 0.03 7.45 – 0.005 7.47 – 0.02

HCO3- 10 min 25.88 – 0.9 25.78 – 1.6 22.58 – 1.3 26.93 – 1.3

210 min 22.36 – 1.9 19.76 – 2.4 24.48 – 0.5 25.98 – 1

*Denotes significance ( p < 0.05) between an experimental group and sham animals at that time point (**p < 0.01; ***p < 0.001).{Denotes significance ( p < 0.05) between time points within an experimental group ({{p < 0.01; {{{p < 0.001).

FIG. 1. Changes in rat mean arterial blood pressure (MABP) following traumatic brain injury (TBI) induced by lateral fluidpercussion (LFP) injury (A) or impact-acceleration weight drop (WD) (B). Significant ( p < 0.05) hypotension relative to shamanimals developed with TBI alone. - = TBI alone (LFP n = 4; WD n = 6); , = TBI plus hypoxia (LFP n = 6; WD n = 6);�= hypoxia alone (LFP n = 4; WD n = 4); B = shams (LFP n = 4; WD n = 5).

2106 GABRIELIAN ET AL.

increased in this group when compared to animals with nointracranial bleeding (Fig. 3). Indeed, ICP approached 20 mmHg by the conclusion of the 4 h monitoring period. Moreover,the scale of the increase in ICP varied and seemed related tothe level of intracranial bleeding. Therefore, the presence of amass lesion facilitated the development of increased ICP inrats following TBI.

CPP analysis

With respect to CPP, sham group animals demonstratedpressures of 100 mm Hg over the 4 h period post-injury(Fig. 4). Following both FPI and WD injury alone, there was asignificant decline in CPP of > 20 mm Hg in both injurygroups relative to shams ( p < 0.05). Surprisingly, the decline inCPP between 1 and 4 h post-trauma was attenuated by theearlier exposure to secondary hypoxia ( p > 0.05 versusshams). The increased CPP in animals previously exposed tohypoxia was most likely related to the effects of MABP to aprior hypoxic episode, given that the CPP changes closelyresembled changes in MABP. Whether this increase in MABPafter the hypoxic episode is a rebound effect in an attempt torestore tissue oxygenation is unknown.

The relationship between MABP and CPP was even moreevident during the 15 min hypoxic period. Given that the ICPprobe could be simply inserted through the existing craniot-omy in FPI animals, very early ICP/CPP measurements couldbe obtained in this subgroup of animals (Fig. 5). During thehypoxic episode immediately after TBI, ICP did not changesignificantly at any stage (mean = 10 – 1 mm Hg). In contrast,CPP decreased from 100 – 3 mm Hg to 42 – 6 mm Hg (Fig. 5A).This profound decrease in CPP was associated with a decreasein MABP during the hypoxic episode (see Fig. 1). CPP stayedlow throughout the entire period of secondary hypoxia (Fig.5B), but when the animals were returned to a normoxic gasmixture, CPP rapidly returned to normal values, in associa-tion with the increase in MABP (see Fig. 1). Such falls in CPPand MABP were not observed in trauma-only animals, sug-

gesting that any decline in MABP during periods of hypoxiaare highly deleterious to CPP.

Discussion

The current study examines ICP changes in two differentand well-characterized models of TBI in the rat: the focal fluidpercussion injury model and the diffuse impact-accelerationmodel, with a particular focus on the effects of secondaryhypoxia and mass lesions on post-traumatic ICP. Our results

FIG. 2. Changes in rat intracranial pressure (ICP) following traumatic brain injury (TBI) induced by lateral fluid percussioninjury (A) or impact-acceleration weight drop (B). ICP in animals subjected to TBI plus hypoxia following impact accelerationinjury (B) was statistically greater than that in sham animals after 105 min (0.001 < p < 0.05), although no significant differenceswere observed between hypoxia only animals and any other group. - = TBI alone; , = TBI plus hypoxia;�= hypoxia alone;B = shams.

FIG. 3. Effects of mass lesions on rat intracranial pressure(ICP) following fluid percussion induced traumatic braininjury (TBI). ICP was significantly greater in TBI animalswith mass lesions than in all other without an intracranialhemorrhage. * = p < 0.05; ** = p < 0.01; *** = p < 0.001 versus TBIalone. : = TBI with intracranial hemorrhage present (n = 5);- = TBI alone (n = 4); , = TBI plus hypoxia (n = 6);� = hypoxia alone (n = 4); B = shams (n = 4).

ICP CHANGES IN RATS AFTER TBI 2107

demonstrate that TBI alone or in combination with secondaryhypoxia does not result in any sustained increase in ICP in therat unless there is a mass lesion associated with a hemorrhage.In the absence of ICP changes, reductions in CPP that werenoted in TBI were closely associated with changes in MABP.

Apnea is frequently observed in rats following experi-mental TBI, however hypoxemia is typically avoided by rapidventilation (Prins and Hovda, 1998). To simulate the effect ofpost-traumatic apnea in these studies, animals received a15 min hypoxic episode immediately following TBI. Thehypoxic episode reduced arterial PO2 to * 50 mm Hg. Thisfall in oxygen saturation to 78.4 – 1.2 % would correlate to asignificant increase in mortality and morbidity for human TBIsurvivors (Stocchetti et al., 1996).

Animals exposed to secondary hypoxia in addition to TBIshowed some increase in ICP compared to the sham groups,albeit that this increase was very modest and never achievedstatistical significance in the LFP group. Statistical significancewas achieved in the WD model when compared to sham

animals; however again, this increase of ICP was modest(2–4 mm Hg over hypoxia alone) and would not be consid-ered clinically relevant. Nonetheless, the increase in the WDmodel compared to the LFP model may reflect the differentialnature of edema formation in the two models of TBI. Earlyedema is typically considered vasogenic in the WDI model,although an element of cytotoxic edema is anticipated becauseof delayed secondary injury factors, as previously discussed(Unterberg et al., 2004). However, introduction of a hypoxicinsult to this model may have resulted in a more rapid andsubstantial cytotoxic edema which, combined with the vaso-genic edema, may have resulted in an increase in the brainwater content over and above the 2–4% typically reported(O’Connor et al., 2006). This is in contrast with LFP wherecytotoxic edema has been reported to dominate in the earlyphase after trauma (Albensi et al., 2000; Bareyre et al., 1997).

In both models of TBI, cytotoxic swelling of neurons andglia does not occur in isolation but in conjunction with vaso-genic accumulation of fluid in the extracellular space.

FIG. 4. Changes in rat cerebral perfusion pressure (CPP) following traumatic brain injury (TBI) induced by lateral fluidpercussion injury (A) or impact-acceleration weight drop (B). A significant ( p < 0.05) decline in CPP developed with TBI alonerelative to sham animals. - = TBI alone; , = TBI plus hypoxia; �= hypoxia alone; B = shams.

FIG. 5. Changes in rat cerebral perfusion pressure (CPP) and intracranial pressure (ICP) during and following the 15 minhypoxic episode induced immediately after FPI. (A) Shows group data from all animals. (B) Shows a representative raw datatrace of rat CPP from a single animal. The spike at 0 represents the beginning of the hypoxic period. CPP was significantlydecreased throughout the hypoxic episode before recovering to normal values after animals were returned to normoxia.

2108 GABRIELIAN ET AL.

Intracellular swelling is therefore not balanced by a decreasein extracellular fluid. On the contrary, intracellular expansionwill occur in concert with expansion of the extracellular space,ultimately resulting in substantial enlargement of cerebraltissue. The extracellular space only constitutes 12–19% ofbrain volume (Go, 1997), therefore swelling of the intracellularspace along with the introduction of new intravascular con-stituents (i.e., water and plasma solutes) provides greaterpotential for tissue expansion than does vasogenic edema inisolation.

A number of previous studies have reported ICP values inrodents after TBI, with some reporting increases after TBI (Liuet al., 2002; Wei et al., 2010; Zweckberger et al., 2009) andothers reporting no increase (Goren et al., 2001; Kahveci et al.,2001; Rogatsky et al., 2003; Thomas et al., 1998). Notably,those that report increases have usually been associated witha severe focal or penetrating injury that causes extensivehemorrhage. Although the reasons for such contradictions areunclear, the findings may well be explained by anatomicaldifferences in relation to the tentorium cerebelli in the centralnervous system of rats compared to other mammals such ashumans, sheep, pigs, cats, or dogs. The tentorium cerebelli isan arched lamina of dura mater separating the supratentorialcompartment (containing the cerebral hemispheres) from theinfratentorial compartment, or posterior fossa, which containsthe cerebellar hemispheres and brainstem (Klintworth, 1968).In mice, rats, hamsters, and gerbils, it consists of delicatemeningeal folds that only separate the lateral portions of thecerebellum and cerebral hemispheres. In other mammals in-cluding goats, sheep, cats, dogs, and humans, the bilateralfolds have joined to form a diaphragm, a structure sufficientto prevent expansion or collapse of the supratentorial matterinto the infratentorial compartment (Bull, 1969). Such move-ment through the tentorial aperture is a feature of transten-torial herniation. Given that the tentorium cerebelli of the ratis quite rudimentary and underdeveloped, any pressure as-sociated with brain swelling could potentially be readilytransferred to the infratentorial compartment and beyond,thus reducing the supratentorial ICP. No post-traumatic in-crease in ICP would be noted. In the presence of hemorrhage,however, the mass lesion will overcome the volume expan-sion allowed by the absence of the tentorium (reduced com-pliance) and an increase in ICP will result. Therefore theinconsistencies in the post-traumatic ICP reported in rodentsmay well be related to the presence or absence of intracranialhemorrhage. Moreover, the presence or absence of increasedICP would have a critical impact on the development of ef-fective neuroprotective therapies designed for translation intothe human condition where increases in ICP are common.

Conclusions

Following TBI in rats, an elevation of ICP was only ob-served in animals that developed mass lesions. TBI alone orcoupled with secondary hypoxia did not result in any relevantincrease of ICP in both models of TBI in the absence of a masslesion. In the absence of any change in ICP, CPP was driven bychanges in MABP. Although the reasons for the lack of ICPchange are unknown, we speculate that rats may be able tocompensate for the intracranial expansion associated withcerebral edema along their craniospinal axis, partially becauseof an underdeveloped tentorium cerebelli. The development

of neuroprotective treatments targeting the effects of in-creased ICP might therefore be best pursued in other speciesthat may more consistently produced sustained increases inpost-traumatic ICP.

Acknowledgments

We would thank Dr. Emma Thornton for her helpfulcomments during the preparation of this manuscript. Thisresearch was funded, in part, by a grant from the NationalHealth and Medical Research Council of Australia.

Author Disclosure Statement

No competing financial interests exist.

References

Albensi, B.C., Knoblach, S.M., Chew, B.G., O’Reilly, M.P., Faden.A.I., and Pekar, J.J. (2000). Diffusion and high resolution MRIof traumatic brain inury in rats: time course and correlationwith histology. Exp. Neurol. 162, 61–72.

Atkinson, J.L., Anderson, R.E., and Murray, M.J. (1998). Theearly critical phase of severe head injury: importance of apneaand dysfunctional respiration. J. Trauma 45, 941–945.

Bareyre, F., Wahl, F., McIntosh, T.K., and Stutzmann, J.M.(1997). Time course of cerebral edema after traumatic braininjury in rats: effects of riluzole and mannitol. J. Neurotrauma14, 839–49.

Beauchamp, K., Mutlak, H., Smith, W.R., Shohami, E., and Sta-hel, P.F. (2008). Pharmacology of traumatic brain injury:where is the ‘‘Golden Bullet’’? J. Mol. Med. 14, 731–740.

Bruns, J., Jr., and Hauser, W.A., (2003). The epidemiology oftraumatic brain injury: a review. Epilepsia 44, 2–10.

Bull, J.W. (1969) Tentorium cerebelli. Proc. R. Soc. Med. 62, 1301–1310.

Bullock, R.M., Chesnut, R., Clifton, G.L., Ghajar, J., Marion,D.W., Narayan, R.K., Newell, D.W., Pitts, L.H., Rosner, M.J.,Walters, B.C., and Wilberger, J.E. (2000). Management andprognosis of severe traumatic brain injury, Part 1: Guidelinesfor the management of severe traumatic brain injury. J. Neu-rotrauma. 17, 471–478.

Cernak, I. (2005). Animal models of head trauma. Neurother-apeutics 2, 410–422.

Donkin, J.J., and Vink, R. (2010). Mechanisms of cerebral edemain traumatic brain injury: therapeutic developments. Curr.Opin. Neurol. 23, 293–299.

Donkin, J,J., Nimmo, A.J., Cernak, I., Blumbergs, P.C., and Vink,R. (2009) Substance P is associated with the development ofbrain edema and functional deficits after traumatic brain in-jury. J. Cereb. Blood Flow Metab. 29, 1388–1398.

Feickert, H.J., Drommer, S., and Heyer, R. (1999). Severe headinjury in children: impact of risk factors on outcome. J. Trau-ma 47, 33–38.

Finfer, S.R., and Cohen, J. (2001). Severe traumatic brain injury.Resuscitation 48, 77–90.

Fish, R. (1997) Pharmacology of injectable anesthetics, in: An-esthesia and Analgesia in Laboratory Animals. D.F. Kohn, S.K.Wixson, W.J. White, and G.J. Benson (eds.), Academic Press:Sydney, pps. 6–7.

Fleminger, S., and Ponsford, J. (2005). Long term outcome aftertraumatic brain injury. B.M.J. 331, 1419.

Gardenfors, A., Nilsson, F., Skagerberg, G., Ungerstedt, U., andNordstrom, C.-H. (2002). Cerebral physiological and bio-chemical changes during vasogenic brain edema induced by

ICP CHANGES IN RATS AFTER TBI 2109

intrathecal injection of bacterial lipopolysaccharides in piglets.Acta Neurochir. 144, 601–609.

Go, K.G. (1997). The normal and pathological physiology ofbrain water. Adv. Tech. Stand. Neurosurg. 23, 47–142.

Goren, S., Kahveci, N., Alkan, T., Goren, B., and Korfali, E.(2001). The effects of sevoflurane and isoflurane on intracra-nial pressure and cerebral perfusion pressure after diffusebrain injury in rats. J. Neurosurg. Anesthesiol. 13, 113–119.

Harris, C., DiRusso, S., Sullivan, T., and Benzil, D.L. (2003).Mortality risk after head injury increases at 30 years. J. Am.Coll. Surg. 197, 711–716.

Ishige, N., Pitts, L.H., Berry, I., Carlson, S.G., Nishimura, M.C.,Moseley, M.E., and Weinstein, P.R. (1987). The effect of hyp-oxia on traumatic head injury in rats: alterations in neurologicfunction, brain edema, and cerebral blood flow. J. Cereb.Blood Flow Metab. 7, 759–767.

Katayama, Y., and Kawamata, T. (2003). Edema fluid accumu-lation within necrotic brain tissue as a cause of the mass effectof cerebral contusion in head trauma patients. Acta Neurochir.86, 323–327.

Kahveci, F.S., Kahveci, N., Alkan, T., Goren, B., Korfali, E., andOzluk, K. (2001). Propofol versus isoflurane anesthesia underhypothermic conditions: effects on intracranial pressure andlocal cerebral blood flow after diffuse traumatic brain injury inthe rat. Surg. Neurol. 56, 206–214.

Kemp, A.M., Stoodley, N., Cobley, C., Coles, L., and Kemp,K.W. (2003). Apnoea and brain swelling in non-accidentalhead injury. Arch. Dis. Child. 88, 472–476.

Klatzo, I. (1967). Presidental address. Neuropathological aspectsof brain edema. J. Neuropath. Exp. Neurol. 26, 1–14.

Klintworth, G.K. (1968) The comparative anatomy and phylog-eny of the tentorium cerebelli. Anat. Rec. 160, 635–642.

Langlois, J.A., Rutland-Brown, W., and Wald, M.M. (2006). Theepidemiology and impact of traumatic brain injury: a briefoverview. J. Head Trauma Rehabil. 21, 375–378.

Liang, D., Bhatta, S., Gerzanich, V., and Simard, J.M. (2007).Cytotoxic edema: mechanisms of pathological cell swelling.Neurosurg. Focus 22, 1–9.

Lighthall, J.W., Dixon, C.E., and Anderson, T.E. (1989). Experi-mental models of brain injury. J. Neurotrauma 6, 83–97.

Liu, H., Goodman, J.C., and Robertson, C.S. (2002). The effects ofL-arginine on cerebral hemodynamics after controlled corticalimpact injury in the mouse. J. Neurotrauma 19, 327–334.

Maggi, C.A., and Meli, A. (1986) Suitability of urethane anes-thesia for physiopharmacological investigations in varioussystems. Part 1. General considerations. Experientia 42, 109–114.

Marmarou, A. (1992). Increased intracranial pressure in head injuryand influence of blood volume. J. Neurotrauma. 9, S327–332.

Marmarou A. (2007). A review of progress in understanding thepathophysiology and treatment of brain edema. Neurosurg.Focus. 22, 1–10.

Marmarou, A., Anderson, R.L., Ward, J.D., Choi, S.C., Young,H.F., Eisenberg, H.M., Foulkes, M.A., Marshall, L.F., and Jane,J.A. (1991). Impact of ICP instability and hypotension onoutcome in patients with severe head trauma. J. Neurosurg.75, S59–66.

Marmarou, A., Foda, M.A., van den Brink, W., Campbell, J.,Kita, H., and Demetriadou, K. (1994). A new model of diffusebrain injury in rats. Part I: Pathophysiology and biomechanics.J. Neurosurg. 80, 291–300.

McIntosh, T.K., Soares, H., Thomas, M., and Cloherty, K.(1990). Development of regional cerebral oedema after lateralfluid-percussion brain injury in the rat. Acta Neurochir. 51,263–264.

McIntosh, T.K., Vink, R., Noble, L., Yamakami, I., Fernyak, S.,Soares, H., and Faden, A.L. (1989). Traumatic brain injury inthe rat: characterization of a lateral fluid-percussion model.Neuroscience 28, 233–244.

Miller, J.D., Becker, D.P., Ward, J.D., Sullivan, H.G., Adams,W.E., and Rosner, M.J. (1977). Significance of intracranialhypertension in severe head injury. J. Neurosurg. 47, 503–516.

Nida, T.Y., Biros, M.H., Pheley, A.M., Bergman, T.A., andRockswold, G.L. (1995) Effect of hypoxia or hyperbaric oxygenon cerebral edema following moderate fluid percussion orcortical impact injury in rats. J. Neurotrauma 12, 77–85.

O’Connor, C.A., Cernak, I., and Vink, R. (2006). The temporalprofile of edema formation differs between male and femalerats following diffuse traumatic brain injury. Acta Neurochir.Suppl. 96, 121–124.

Papadopoulos, M.C., Manley, G.T., Krishna, S., and Verkman,A.S. (2004). Aquaporin-4 facilitates reabsorption of excessfluid in vasogenic brain edema. FASEB J. 10, 1–18.

Prins, M.L., and Hovda, D.A. (1998). Traumatic brain injury inthe developing rat: effects of maturation on Morris watermaze acquisition. J. Neurotrauma 15, 799–811.

Rogatsky, G.G., Mayevsky, A., Zarchin, N., and Doron, A.(1996). Continuous multiparametric monitoring of brain ac-tivities following fluid-percussion injury in rats: preliminaryresults. J. Basic Clin. Physiol. Pharmacol. 7, 23–43.

Rogatsky, G.G., Sonn, J., Kamenir, Y., Zarchin, N., andMayevsky, A. (2003). Relationship between intracranialpressure and cortical spreading depression following fluidpercussion brain injury in rats. J. Neurotrauma 20, 1315–1325.

Rolett, E.L., Azzawi, A., Liu, K.J., Yongbi, M.N., Swarts, H.M.,and Dunn, J.F. (2000). Critical oxygen tension in rat brain: acombined (31)P-NMR and EPR oximetry study. Am. J. Phy-siol. Regul. Integr Comp. Physiol. 279, R9–R16.

Schilling, L., and Wahl, M. (1999). Mediators of cerebral edema.Adv. Exp. Med. Biol. 474, 123–141.

Schneider, G.H., Baethmann, A., and Kempski, O. (1992). Me-chanisms of glial swelling induced by glutamate. Can. J.Physiol. Pharmacol. 70, 334–343.

Soares, H.D., Thomas, M., Cloherty, K., and McIntosh, T.K.(1992). Development of prolonged focal cerebral edema andregional cation changes following experimental brain injury inthe rat. J. Neurochem. 58, 1845–1852.

Soustiel, J.F., and Larisch, S. (2010). Mitochondrial damage: atarget for new therapeutic horizons. Neurotherapeutics 7, 13–21.

Stahl, N., Ungerstedt, U., and Nordstrom, C.H. (2001). Brainenergy metabolism during controlled reduction of cerebralperfusion pressure in severe head injuries. Intensive CareMed. 27, 1215– 1223.

Steiner, L.A., and Andrews, P.J.D. (2006). Monitoring the injuredbrain: ICP and CBF. Br. J. Anaesth. 97, 26–38.

Stocchetti, N., Furlan, A., and Volta, F. (1996). Hypoxemia andarterial hypotension at the accident scene in head injury. J.Trauma 40, 764–767.

Thomas, S., Tabibnia, F., Schuhmann, M.U., Hans, V.H., Brinker,T., and Samii, M. (1998). Traumatic brain injury in the devel-oping rat pup: studies of ICP, PVI and neurological response.Acta Neurochir. Suppl. 71, 135–137.

Thompson, H.J., Lifshitz, J., Marklund, N., Grady, M.S., Graham,D.I., Hovda, D.A., and McIntosh, T.K. (2005). Lateral fluidpercussion brain injury: a 15-year review and evaluation. J.Neurotrauma 22, 42–75.

2110 GABRIELIAN ET AL.

Tolias, C.M., and Bullock, M.R. (2004). Critical appraisal ofneuroprotection trials in head injury: what have we learned?Neurotherapeutics 1, 71–79.

Unterberg A.W., Stover, J., Kress, B., and Kiening, K.L. (2004).Edema and brain trauma. Neuroscience 129, 1021–1029.

Van Putten, H.P., Bouwhuis, M.G., Muizelaar, J.P., Lyeth, B.G.,and Berman, R.F. (2005) Diffusion–weighted imaging of ede-ma following traumatic brain injury in rats: effects of sec-ondary hypoxia. J. Neurotrauma 22, 857–872.

Vink, R., and Nimmo, A.J. (2009). Multifunctional drugs forhead injury. Neurotherapeutics. 6, 28–42.

Wahlstrom, M.R., Olivecrona, M., Koskinen, L.-O.D., Rydenhag,B., and Naredi, S. (2005). Severe traumatic brain injury inpediatric patients: treatment and outcome using an intracra-nial pressure targeted therapy – the Lund concept. IntensiveCare Med. 31, 832–839.

Wei, G., Lu, X.C., Yang, X., and Tortella, F.C. (2010). Intracranialpressure following penetrating ballistic-Like brain Injury inrats. J. Neurotrauma 27, 1635–1641.

Zweckberger, K., and Plesnila, N. (2009). Anatibant, a selectivenon-peptide bradykinin B2 receptor antagonist, reduces in-tracranial hypertension and histopathological damage afterexperimental traumatic brain injury. Neurosci Lett. 454, 115–117.

Address correspondence to:Robert Vink, Ph.D.

School of Medical SciencesUniversity of Adelaide

Adelaide SA, 5005Australia

E-mail: Robert.Vink@adelaide.edu.au

ICP CHANGES IN RATS AFTER TBI 2111