Available online at www.sciencedirect.com

Surgical Neurology 71 (2009) 54–59www.surgicalneurology-online.com

Ischemia

The neuroprotective effect of dexmedetomidine in the hippocampus ofrabbits after subarachnoid hemorrhage

Murat Cosar, MDa,⁎, Olcay Eser, MDa, Huseyin Fidan, MDb, Onder Sahin, MDc,Sadik Buyukbas, MDd, Yuksel Ela, MDb, Murat Yagmurca, MDe, Oguz A. Ozen, MDf

aNeurosurgery, bAnesthesiology, and cPathology Department, Faculty of Medicine, Kocatepe University, 03000 Afyonkarahisar, TurkeydBiochemistry Department, Faculty of Medicine, Selcuk University, 42300 Konya, Turkey

eHistology, and fAnatomy Department, Faculty of Medicine, Kocatepe University, 03000 Afyonkarahisar, Turkey

Received 19 March 2007; accepted 7 August 2007

Abstract Background: Subarachnoid hemorrhage is a serious condition, often accompanied by cerebral

Abbreviations: CSand manganez; H andhydrogen peroxide;intravenous; MDA,SAH, subarachnoid heoxidase.

⁎ CorrespondingTel.: +90 505 80413

E-mail address: d

0090-3019/$ – see frodoi:10.1016/j.surneu.2

vasospasm, which may lead to brain ischemia and neurologic deterioration. We evaluated ifdexmedetomidine has neuroprotective effects in the hippocampus of vasospastic SAH rabbits or not.Materials and Methods: Eighteen New Zealand rabbits were taken. An experimental SAH modelwas formed by injecting 0.9 mL of autologous arterial blood per 1 kg of body weight to the cisternamagna of 12 rabbits. Craniotomy was performed in the control group (n = 6) except performingexperimental SAH. Rabbits in the SAH-alone (n = 6) group were infused with 5 mL · kg−1 · h−1

0.9% sodium chloride, and rabbits (n = 6) in the SAH-dexmedetomidine group were infused with5 μg · kg−1 · h−1 dexmedetomidine for 2 hours, 48 hours after SAH was established. Rabbits of allgroups were sacrificed via penthotal 24 hours after dexmedetomidine administration. Brains wereremoved immediately, and hippocampal tissues were blocked from the right hemisphere forhistopathologic study. In addition to this, hippocampal tissues of left hemispheres were dissected forbiochemical analyses to evaluate MDA levels, activity of XO, and SOD.Results: The histopathologic study showed that dexmedetomidine may have a neuroprotective effectin SAH-induced hippocampal injuries. The biochemical parameters support the neuroprotectiveeffect of dexmedetomidine (P b .05).Conclusion: Our study showed that dexmedetomidine may have a neuroprotective effect in thehippocampus of vasospastic SAH rabbits.© 2009 Elsevier Inc. All rights reserved.

Keywords: Brain ischemia; Dexmedetomidine; Neuroprotective; Subarachnoid hemorrhage

1. Introduction

Subarachnoid hemorrhage is characterized by a high levelof morbidity and mortality due to the development ofcerebral vasospasm [17]. The incidence of vasospasm has

F, cerebrospinal fluid; Cu-Zn and Mn, copper-zincE, hemotoxylin-eosine; HCI, hydrogen clorur; H2O2,ICP, intracranial pressure; IM, intramuscular; IV,malondialdehyde; ROS, reactive oxygen species;morrhage; SOD, superoxide dismutase; XO, xanthine

author. P.K: 34, 03000 Afyonkarahisar, Turkey.62 (mobile).rcosar@hotmail.com (M. Cosar).

nt matter © 2009 Elsevier Inc. All rights reserved.007.08.020

been estimated to occur in one third of all patients with SAH[2,28]. Cerebral vasospasm may lead to brain ischemia andneurologic deterioration.

The pathogenesis of cerebral vasospasm is unclear, butblood breakdown products [1,16] and elevated levels ofcirculating catecholamines may be involved in the pathogen-esis behind these events [22]. It has been demonstrated thatvasospasm can be influenced by many factors, among whichthe possible role of the sympathetic nervous system has beensuggested [22].

If the increase in norepinephrine release is a causativecompanent of injury, α2-adrenergic agonists may provideprotection against the damaging effects of cerebralischemia. In support of this, an α2-adrenoceptor agonist,

55M. Cosar et al. / Surgical Neurology 71 (2009) 54–59

dexmedetomidine, has also been shown to reduce ischemicdamage, possibly by attenuating the excessive release ofnoradrenaline [12,21]. There is evidence that the selectiveα2-adrenergic agonist, dexmedetomidine, partially preventsneuronal damage in some models of cerebral ischemia. Forinstance, it has been reported that it protects brain tissueagainst incomplete ischemia in rats [9] and against focalischemia in rabbits [20]. Neuroprotection of dexmedetomi-dine is also shown in a gerbil global ischemia model [15].Ma et al [19] declared that the neuroprotective effect ofdexmedetomidine is mediated by activation of the α2A-adrenergic receptor subtype. Our previous studies showedthat dexmedetomidine may decrease the severity ofvasospasm in the basilar artery and has a neuroprotectiveeffect in the prefrontal cortex of SAH rabbits [6].

In this study, we aimed to show a neuroprotective effect ofdexmedetomidine in the hippocampus after vasospasticSAH, which may lead to ischemia and neural deterioration.We used a rabbit SAH model because of the anatomicalsimilarity of the sympathetic innervation of rabbit cerebralarteries as compared with human ones [23].

2. Materials and methods

2.1. Experimental design

This study was approved by the animal ethics committeeof Afyon Kocatepe University (Afyonkarahisar, Turkey).One day before surgery, rabbits were fasted and pretreatedwith an antibiotic, cefazolin (Cefozin, 50 mg/kg IV; Bilim,Istanbul, Turkey). The study was carried out on eighteen1500- to 2000-g male New Zealand rabbits. Rabbits wereanesthetized by xylazin HCI (15 mg/kg) and ketamine(25 mg/kg) IM and placed on a heated surgical table tomaintain the body temperature of the animal at 37°C. Afteranesthesia, 2 L/min of oxygen was given for rabbits not toexperience hypoxia. Rabbits were separated into 3 groups,randomly. All rabbits were fixed in a prone position, and amidline incision between their ears over their externaloccipital protuberance was made caudally to reach thecisterna magna. After the first incision, dermal and subdermaltissues and fascia and paravertebral muscles were dissected.The atlanto-occipital membrane was dissected, and a27-gauge needle was inserted through the dura and thearachnoid membrane into the cisterna magna. The dura wasnot punctured in the control group (n = 6). Nonheparinizedarterial blood (0.9 mL/kg) that was obtained from the earartery was injected slowly into the cisterna magna of12 rabbits to form vasospasm, except the control group. Topermit a good distribution of the blood around the basalintracranial arteries, rabbits were kept in a head-downposition from the onset of the injection until 10 minutesafter injection. This procedure is similar to the method ofSolomon et al [25]. After rabbits recovered fromanesthesia, they were left for 48 hours to their cages forthe establishment of acute vasospasm. All animals in the

3 groups were awake in 1 hour, and they began to eat anddrink 2 hours after returning to their cages. The bodytemperatures (mean, 36.4°C) and physiologic findings of the3 groups animals were in the normal range during 48 hours.

Forty-eight hours after SAH, all rabbits were reanesthe-tized by xylazin HCI (15 mg/kg) and ketamine (25 mg/kg) IMand placed on a heated surgical table to maintain bodytemperature at 37°C. After anesthesia was introduced, 2 L/minof oxygen was given. Half the dose of anesthetics wasreintroduced to rabbits when needed. All rabbits werereanesthetised when they moved or when their heart beatsincreased by 10%. One of the ear arteries was cannulated tomonitor invasive arterial blood pressure. One of the ear veinswas cannulated for IV fluid infusion. Electrocardiographicreadings, invasive blood pressure levels, peripheral oxygensaturation, and temperature were monitored (Datex OhmedaType F-CU8; Datex Instrumentarium, Helsinki, Finland)throughout the procedure. The control group received5 mL · kg−1 · h−1 of 0.9% sodium chloride IV for 2 hours.Six of the rabbits in the SAH-alone group received 5 mL ·kg−1 · h−1 of 0.9% sodium chloride via IV infusion for 2hours, 48 hours from the vasospastic process. Six of therabbits in the SAH-dexmedetomidine group received dexme-detomidine 5 μg · kg−1 · h−1 via IV infusion for 2 hours, 48hours from the vasospastic process. Intracerebral pressure(mean, 7 mm Hg) and pericranial temperature (mean, 36.6°C)were measured in all groups during these processes (Camino110-4BT; Integra NeuroCare, San Diego, Calif).

The members of all groups were killed with 20 mg/kgpenthotal after 24 hours of drug receiving. Brains wereremoved from the skull immediately, and hippocampal tissueof the right hemispheres were blocked for histopathologicstudy. The hippocampal tissues of left hemispheres weredissected, including the hippocampal formation for theevaluation of biochemical analyses; so, the MDA levels,activity of XO, and SOD were measured from the lefthippocampal tissues.

2.2. Histopathologic procedures

The right hemisphere samples including the hippocampuswere fixed in 10% neutral buffered formalin and stored at4°C. One week after the sampling, the samples wereremoved and placed in fresh fixative. Fixed tissue sampleswere processed routinely by paraffin embedding technique.The coronal hippocampal tissue sections 5 μm in thicknesswere cut with a Leica RM 2145 microtome (LeicaInstruments GmbH, Nussloch, Germany) at 200-μm inter-vals, and every eighth section through the brain wascollected on slide and stained with H and E. Preparationswere evaluated by a bright field microscope (Eclipse E600Wmicroscope; Nikon, Tokyo, Japan) and were photographedwith a microscope digital camera (Nikon DP70; Nikon).

The numbers of neurons in hippocampal CA1 region werecalculated by an image analysis system. The investigatorwho performed these measurements was unaware of the

56 M. Cosar et al. / Surgical Neurology 71 (2009) 54–59

experiment. The system used is composed of a personalcomputer, hardware and software camera (the images wereprocessed by an IBM-compatible personal computer, high-resolution video monitor, and image analysis software[BS200Doc, version 2.0; BAB Imaging Systems, Ankara,Turkey]), and optical microscope. The method requirespreliminary software procedures of spatial calibration (micronscale) and setting of color segmentation for quantitativecolor analysis.

2.3. Assay of hippocampal tissue MDA

Malondialdehyde levels were estimated by the double-heatingmethod of Draper andHadley [4]. The principle of themethod is the spectrophotometric measurement of the colorgenerated by the reaction of thiobarbituric acid with MDA.For this purpose, 2.5 mL of 100 g/L trichloroacetic acidsolution was added to 0.5 mL of sample (left hippocampaltissues supernatant) in each centrifuge tube, and the tubeswere placed in a boiling water bath for 15 minutes. Aftercooling in tap water, the tubes were centrifuged at 1000g for10 minutes; 2 mL of the supernatant was then added to 1 mLof 6.7 g/L thiobarbituric acid solution in a test tube, and thetube was placed in a boiling water bath for 15 minutes. Thesolution was then cooled in tap water, and its absorbance wasmeasured using a spectrophotometer (Shimadzu UV-1601;Shimadzu, Tokyo, Japan) at 532 nm. Hippocampus MDAlevels were also determined by the absorbance coefficient ofthe MDA-thiobarbituric acid complex and were expressed asnmol/g wet tissue.

2.4. Assay of hippocampal tissue XO activity

Xanthine oxidase activity was measured spectrophoto-metrically by the formation of uric acid from xanthinethrough the increase in absorbance at 293 nm [24]. One unitof activity was defined as 1 mmol uric acid formed perminute at 37°C and pH of 7.5. Results were expressed inunits per gram of protein for hippocampal tissues.

Fig. 1. Histopathologic section of hippocampus CA1 region in different groups. Achanges and shrinkage in cytoplasm and extensively dark picnotic nuclei are seen inand slight shrinking in cytoplasm and nuclei are seen in neuronal cells. Scale bar

2.5. Assay of hippocampal tissue SOD activity

Total (Cu-Zn and Mn) SOD activity was determinedaccording to the method of Sun et al [27]) with a slightmodification by Durak et al [5]. The principle of the methodis based briefly on the inhibition of nitroblue tetrazoliumreduction by the xanthine/XOD system as a superoxidegenerator. Activity was assessed in the ethanol phase ofthe supernatant after 1.0 mL ethanol/chloroform mixture(5/3, v/v) was added to the same volume of sample(hippocampal tissue supernatant) and centrifuged. One unitof SOD was defined as the enzyme amount causing 50%inhibition in the nitroblue tetrazolium reduction rate. Super-oxide dismutase activity was expressed as units permilligram of protein for hippocampal tissues.

2.6. Statistical analyses

All statistical analyses were carried out using SPSSstatistical software (SPSS for windows, version 11.0; SPSS,Chicago, Ill). All data are presented as mean ± SD. Differencesin measured histopathologic and biochemical parametersamong the 3 groups were considered as nonparametric andanalyzed by Kruskal-Wallis H test and Mann-Whitney U test.P b .05 was considered statistically significant.

3. Results

Histopathologic results of the present study showed thatdexmedetomidine was able to offer neuroprotection againstSAH-induced hippocampal injuries. When the histopathol-ogy of the hippocampus was studied, the SAH-dexmedeto-midine group showed decline of neural injury findings.

In addition, the morphology of neurons in the CA1 regionof the hippocampal tissue was normal in the control group(Fig. 1A). The most consistent findings in histologic sectionsof hippocampal CA1 regions in SAH group stained withH and E were severe degenerative changes, shrunkencytoplasm, and extensively dark picnotic nucleus (Fig. 1B).

: Appearance in control rabbit. B: SAH-group rabbit. Severe degenerativeneuronal cells. C: SAH-dexmedetomidine rabbit. Less degenerative changes= 25 μm.

Fig. 2. The numbers (neuron/mm2) of neurons in the hippocampal CA1

region of control, SAH, and SAH-dexmedetomidine groups. The data areexpressed as mean ± SD (n = 6).

57M. Cosar et al. / Surgical Neurology 71 (2009) 54–59

In the SAH-dexmedetomidine group, the severity ofdegenerative changes in the cytoplasma and especiallynucleus of hippocampal tissues were less than in theSAH group (Fig. 1C). The number of neurons in thehippocampal CA1 region for the SAH group was alsosignificantly less than for the control and SAH-dexmedeto-midine groups (P b .05) (Fig. 2).

Malondialdehyde levels and activity of XO and SOD ofhippocampal tissue were evaluated. The MDA levels weresignificantly higher in the SAH-alone group than in thecontrol and SAH-dexmedetomidine groups (P b .05).Xanthine oxidase activity was significantly high in theSAH-alone group than in the control and SAH-dexmede-tomidine groups (P b .05). Superoxide dismutase activitywas significantly lower in the SAH-alone group than inthe control and SAH-dexmedetomidine groups (P b .05)(Table 1).

There was no difference between the groups inelectrocardiographic results, intracerebral pressure, invasiveblood pressure, peripheral oxygen saturation, and pericra-nial and body temperature levels measured during theexperimental study.

Table 1Oxidant and antioxidant status of control, SAH-alone, and SAH-dexmedetomidine groups in the hippocampal formation of rabbits

Control(n = 6)

SAH (n = 6) SAH-dexmedetomidine(n = 6)

MDA, wet tissue(nmol/g)

15.20 ± 2.31 21.66 ± 3.06 ⁎ 16.92 ± 2.62 †

XO, protein (U/g) 0.29 ± 0.10 0.51 ± 0.08 ⁎ 0.27 ± 0.06 †

SOD, protein(U/mg)

0.13 ± 0.02 0.08 ± 0.01 ⁎ 0.12 ± 0.03

Data expressed as mean ± SD.⁎ P b .05, compared with control.† P b .05, compared with SAH.

4. Discussion

We evaluated the neuroprotective effects of dexmedeto-midine in hippocampal tissue of experimental SAH. Themodel was formed via autologous blood injection into thecisterna magna of rabbits. The results of the study providedthat a selective α2 agonist, dexmedetomidine, may have aneuroprotective effect in the hippocampal tissue of vasos-pastic SAH rabbits.

Although it is known that vasospasm after SAH is amultifactorial situation, previous clinical and experimentalstudies showed that there is a marked activation of thesympathetic nervous system in vasospastic process of SAH

[17]. The sympathetic stimulus that induces the formation ofnorepinephrine also causes abnormal sensitivity of thecerebral vasculature to increased catecholamines on arterialwalls. The cerebral vascular system may also be degeneratedbecause of the increased catecholamine levels [18].Increased catecholamine levels after SAH may lead tocerebral vasospasm, and this process leads to cerebralischemia. The hippocampus is one of the most vulnerabletissues for ischemia in the brain. There is a massive increasein extracellular norepinephrine in the hippocampus duringischemic insult [7]. The neuroprotective mechanisms ofdexmedetomidine in the prevention of cell death couldinvolve these catecholamines. Dexmedetomidine activatespresynaptic α2-adrenoceptors and could thereby reduceischemic damage by inhibiting catecholamine release andalso vasospasm [11]. Several studies showed that dexmede-tomidine clearly provides neuroprotection against immediatenecrotic cell death and delays the development of infarctionin incomplete ischemia [9,20].

Histopathologic results of our study showed thatdexmedetomidine may have neuroprotective effects againstSAH-induced hippocampal injuries. The severity of degen-erative changes in the cytoplasm and especially nucleus ofhippocampal tissues in the SAH-dexmedetomidine groupwere less than in the SAH-alone group.

Malondialdehyde, XO, and SOD indicators were used toevaluate the oxidative stress and ischemia in hippocampaltissue after vasospastic SAH. The accumulations ofneutrophils in the ischemic tissue after vasospasm are apotential source of ROS [8,30]. Reactive oxygen species arebelieved to cause lipid peroxidation, resulting in damage tobiologic membranes. Xanthine oxidase is a major potentialsource of oxygen free radicals [10]. Cells have a complete setof antioxidant mechanisms that defend the tissues against thedamaging effects of free radicals. It is difficult to quantitatethe ROS because of their reactive nature and short lives; thus,the measurement of MDA (end product of lipid peroxida-tion) and XO activity can be used to estimate the extent ofoxidative stress in tissues [16,26].

The increase in MDA levels and XO activity (in the SAH-alone group) showed that lipid peroxidation induces the

58 M. Cosar et al. / Surgical Neurology 71 (2009) 54–59

production of free radicals, such as hydroxyl radicals.Hydroxyl radicals stimulate the destruction of membranesby setting off a free radical chain reaction known as lipidperoxidation [14]. In our study, MDA levels and XO activityin the vasospastic SAH group were significantly higher thanin the control and SAH-dexmedetomidine groups (P b .05).Our results suggested that dexmedetomidine reduces MDAlevels and XO activity. These results made us think thatdexmedetomidine may have a neuroprotective effect viapreventing ischemia in hippocampal tissue.

Cells have a complete set of antioxidant mechanisms thatdefend the tissues against the damaging effects of freeradicals. Superoxide dismutase is one of the responsiblefactors for the control of superoxide radicals, which iteliminates by converting them to H2O2 [14]. The decrease inthe activity of SOD could be due to increased reactiveoxygen free radicals. The rabbits pretreated with dexmede-tomidine showed increased SOD activity and decreased XOactivity in hippocampal tissue in our study. Our resultssuggest that dexmedetomidine may have the ability toprevent the deleterious effects induced by free radicals viareducing postischemic ROS production by blockingxanthine/XOD. Therefore, the decreased MDA levels andXO activity and elevated SOD activity in the SAH-dexmedetomidine group may show the neuroprotectiveeffects of dexmedetomidine.

In addition to the neuroprotective effects of dexmedeto-midine declared, we must add that the central α2 blockagealleviates potential detrimental effects of metabolizingexcessive noradrenaline, which can lead to the formationof free radicals. As we declared, free radicals are known tocause lipid peroxidation. Although it is reported that CSFlipid hydroperoxide content is related to cerebral vasospasm[3,13], inhibition of oxidative deamination of catechola-mines decreases H2O2 production [29].

Our study points out the benefits of dexmedetomidine invasospastic SAH. Although prevention of SAH is achievedin rabbits by clonidine, another α2 agonist, by starting toestablish it as SAH is formed, this is not the case in mostclinical circumstances. Mostly, patients are admitted byphysicians after they have experienced SAH in differentneurologic status. Our study shows that dexmedetomidinemay be effective for alleviating ischemia and neurologicdeterioration after vasospastic SAH.

Although we have studied the neuroprotective effect ofdexmedetomidine 48 hours after experimental SAH, effec-tiveness of dexmedetomidine after chronic vasospasm mayalso be studied. We think that dexmedetomidine may be usedfor its neuroprotective effects after SAH as well. However,human studies should be done to prove neuroprotectiveeffects of dexmedetomidine and its effects on outcome ofpatients with SAH.

As a result, the major finding of the present studyindicates that dexmedetomidine may have neuroprotectiveeffects 48 hours after acute vasospastic SAH. However,detailed studies are necessary to examine these preliminary

findings and to clarify the mechanisms of action. In thefollowing studies, we are planning to evaluate the effects ofdexmedetomidine on the catecholamine levels, survival, andneurologic complications after SAH.

References

[1] Aydın IH, Onder A. The effect of very early cisternal irrigation onbasilar artery spasm SAH in rat model. Acta Neurochir (Wien) 1991;113:69-73.

[2] Awad IA, Carter LP, Spetzler RF, Medina M, Williams FC. Clinicalvasospasm after subarachnoid hemorrhage: response to hypervolemichemodilution and arterial hypertension. Stroke 1987;18:365-72.

[3] Bunc G, Kovacic S, Strnad S. Sympathetic nervous system exclusionfollowing experimental subarachnoid haemorrhage prevents vasos-pasm in rabbits. Wien Klin Wochenschr 2000;112:533-9.

[4] Draper HH, Hadley M. Malondialdehyde determination as index oflipid peroxidation. Methods Enzymol 1990;186:421-31.

[5] Durak I, Yurtarslani Z, Canbolat O, Akyol O. A methodologicalapproach to superoxide dismutase (SOD) activity assay based oninhibition of nitroblue tetrazolium (NBT) reduction. Clin Chim Acta1993;214:103-4.

[6] Eser O, Cosar M, Fidan H, Sahin O, Buyukbas S, Ela Y, Songur A. Theeffect of dexmedetomidine in the prefrontal cortex of rabbits aftersubarachnoidal hemorrhage. Neurol Psychiatry Brain Res 2006; 13:189-94.

[7] Globus MY, Busto R, Dietrich WD, Martinez E, Ginsberg MD.Intra-ischemic extracellular release of dopamine and glutamate isassociated with striatal vulnerability to ischemia. Neurosci Lett1988;91:36-40.

[8] Granger DN, Korthuis RJ. Physiologic mechanisms of postischemictissue injury. Annu Rev Physiol 1995;57:311-32.

[9] Hoffman WE, Kochs E, Werner C. Dexmedetomidine improvesneurological outcome from incomplete ischemia in the rat. Anesthe-siology 1991;75:328-32.

[10] Ilhan A, Aladag MA, Kocer A, Boluk A, Gurel A, Armutcu F.Erdosteine ameliorates PTZ-induced oxidative stress in mice seizuremodel. Brain Res Bull 2005;65:495-9.

[11] Johansson P, Ehrenstrom F. Presynaptic alpha-adrenoceptor regulationof transmitter release in the conscious rat. Comp Biochem Physiol1988;89:65-9.

[12] Jolkkonen J, Puurunen K, Koistinaho J, Kauppinen R, Haapalinna A,Nieminen L, Sivenius J. Neuroprotection by the α2-adrenoceptoragonist, dexmedetomidine, in rat focal cerebral ischemia. Eur JPharmacol 1999;372:31-6.

[13] Kajita Y, Takayasu M, Dietrich HH. Possible role of nitric oxide inautoregulatory response in rat intracerebral arterioles. Neurosurgery1998;42:834-41.

[14] Kanko M, Maral H, Akbas MH. Protective effects of clopidogrel onoxidant damage in a rat model of acute ischemia. Tohoku J Exp Med2005;205:133-9.

[15] Kuhmonen J, Pokorny J, Miettinen R. Neuroprotective effects ofdexmedetomidine in the gerbil hippocampus following transient globalischemia. Anesthesiology 1997;87:371-7.

[16] Kemaloglu S, Ozkan U, Yilmaz F, Ak E, Acemoglu H, Olmez G.Preventive effects of intracisternal alpha-tocopherol on cerebralvasospasm in experimental subarachnoid haemorrhage. Yonsei MedJ 2003;44:955-60.

[17] Lambert G, Naredi S, Edin E. Sympathetic nervous activationfollowing subarachnoid hemorrhage: influence of intravenous cloni-dine. Acta Anaesthesiol Scand 2002;46:160-5.

[18] Lobato RD, Marin J, Salaices M. Cerebrovascular reactivity tonoradrenaline and serotonin following experimental subarachnoidhemorrhage. J Neurosurg 1980;53:480-5.

59M. Cosar et al. / Surgical Neurology 71 (2009) 54–59

[19] Ma D, Hossain M, Rajakumaraswamy N, Arshad M, Sanders RD,Franks NP, Maze M. Dexmedetomidine produces its neuroprotectiveeffect via the alpha 2A-adrenoceptor subtype. Eur J Pharmacol 2004;502:87-97.

[20] Maeir C, Steinberg GK, Sun GH. Neuroprotection by the α-2adrenoceptor agonist dexmedetomidine in focal model of cerebralischemia. Anesthesiology 1993;79:1-7.

[21] Matsumoto M, ZornowMH, Rabin BC, Maze M. The α2-adrenoceptoragonist, dexmedetomidine, selectively attenuates ischemia-inducedincreases in striatal norepinephrine concentrations. Brain Res1993;627:325-9.

[22] Naredi S, Lambert G, Eden E. Increased sympathetic nervous activityin patients with nontraumatic subarachnoid hemorrhage. Stroke 2000;31:901-6.

[23] Peerless SJ, Yasargil MG. Adrenergic innervation of the cerebral bloodvessels in the rabbit. J Neurosurg 1971;35:148-54.

[24] Prajda N, Weber G. Malignant transformation-linked imbalance:decreased xanthine oxidase activity in hepatomas. FEBS Lett 1975;59(2):245-9.

[25] Solomon RA, Antunes JL, Chen RYZ. Decrease in cerebral blood flowin rats after experimental subarachnoid hemorrhage: a new model.Stroke 1985;16:58-64.

[26] Songur A, Sarsilmaz M, Sogut S, Ozyurt B, Ozyurt H, Zararsiz I,Turkoglu AO. Hypothalamic superoxide dismutase, xantine oxi-dase, nitric oxide, and malondialdehyde in rats fed with fish ω-3fatty acids. Prog Neuro-psychopharmacol Biol Psychiatry 2004;28:693-8.

[27] Sun Y, Oberley LW, Li Y. A simple method for clinical assay ofsuperoxide dismutase. Clin Chem 1988;34:497-500.

[28] Suzuki K, Sarti C, Tuomilehto J. Stroke incidence and case fatality inFinland and in Akita, Japan: a comparative study. Neuroepidemiology1994;13:236-44.

[29] Suzuki T, Akaike N, Ueno KI. MAO inhibitors, clorgyline andlazabemide, prevent hydroxyl radical generation caused by brainischemia/reperfusion in mice. Pharmacology 1995;50:357-62.

[30] Zimmerman BJ, Grisham MB, Granger DN. Role of oxidants inischemia/reperfusion-induced granulocyte infiltration. Am J Physiol1990;258:185-90.

Commentary

The authors report that the selective α2 antagonist,dexmedetomidine, decreases neuronal death in the hippo-campus of rabbits after SAH. Coupled with the evidencesupporting neuroprotective effects of this drug in ischemia, itis worth considering further trials of this agent in SAHmodels. The question is, what model best recreates thehuman condition? Megyesi and colleagues [1] reviewedmodels of vasospasm, but it is apparent that these are largelydesigned to create vasospasm and not the SAH/vasospasmcomplex. The rat endovascular perforation model has beengaining acceptance because it creates SAH more like thatoccurring in humans, but it has important shortcomings thatinclude high variability, production of cerebral ischemiaacutely along with the SAH, and lack of severe, delayedvasospasm. Likely, a better model is needed, but the questionis, who will fund it? I am pessimistic that the National

Institutes of Health could do this. Industry, privatephilanthropy, and the vast wealth of private individualsflying around the world in their private jets need to be tappedas alternate funding sources.

Some factors limit the interpretation of the study by Cosaret al. The control animals did not undergo injection of salineor artificial CSF into the subarachnoid space. The intracra-nial pressure was not measured in the SAH animals, so it ispossible that intracranial pressure changes caused some ofthe neuronal death in the hippocampus. If this was the case,then there would be 2 possible causes for the observedchanges—SAH or increased intracranial pressure. Intracra-nial pressure was monitored, but the article reads like it wasmonitored 48 hours after SAH when the drug injections wereperformed. Also, stereological methods were not used tocount neurons in the hippocampus. Therefore, as mentioned,additional work is needed.

R. Loch Macdonald, MD, PhDDivision of Neurosurgery, St Michael's Hospital

University of Toronto, Toronto, Ontario, Canada M5C 3G7

Reference

[1] Megyesi JF, Vollrath B, Cook DA, Findlay JM. In vivo animal models ofcerebral vasospasm: a review. Neurosurgery 2000;46:448-60.

Response

We appreciate the reviewer's comments and agree withthe statement that we did not use saline for control animals.The reason for this decision was to prevent an increase in thenumber of groups as “control saline” and to minimize thesacrifice of more animals. The ethical committee warnsagainst use of too many animals.

In addition, we used ICP only during the process in allgroups, but we believe that the ICP of the control group mayanswer some of the questions of our study, especially inregard to comparing the groups.

Finally, it would indeed have been better if we had used astereological method for this study, but at this time, it wouldbe difficult because of the selection of our specimens.

Murat Cosar, MDDepartment of Neurosurgery

Faculty of MedicineKocatepe University

Afyonkarahisar, Turkey 03000E-mail address: drcosar@hotmail.com