Xenon Enhances HypothermicNeuroprotection in Asphyxiated

Newborn PigsElavazhagan Chakkarapani, MBBS,1 John Dingley, MD,2 Xun Liu, PhD,1

Nicholas Hoque, MBBS,1 Kristian Aquilina, MD,1

Helen Porter, MD, PhD,3 and Marianne Thoresen, MD, PhD1

Objective: To investigate whether inhaling 50% xenon during hypothermia (HT) offers better neuroprotection thanxenon or HT alone.Methods: Ninety-eight newborn pigs underwent a 45-minute global hypoxic-ischemic insult severe enough tocause permanent brain injury, and 12 pigs underwent sham protocol. Pigs then received intravenous anesthesiaand were randomized to 6 treatment groups: (1) normothermia (NT; rectal temperature 38.5°C, n � 18); (2) 18hours 50% xenon with NT (n � 12); (3) 12 hours HT (rectal temperature 33.5°C, n � 18); (4) 24 hours HT (rectaltemperature 33.5°C, n � 17); (5) 18 hours 50% xenon with 12 hours HT (n � 18); and (6) 18 hours 50% xenonwith 24 hours HT (n � 17). Fifty percent xenon was administered via a closed circle with 30% oxygen and 20%nitrogen. After 10 hours rewarming, cooled pigs remained normothermic until terminal perfusion fixation at 72hours. Global and regional brain neuropathology and clinical neurological scores were performed.Results: Xenon (p � 0.011) and 12 or 24 hours HT (p � 0.003) treatments offered significant histological global,and regional neuroprotection. Combining xenon with HT yielded an additive neuroprotective effect, as there wasno interaction effect (p � 0.54). Combining Xenon with 24 hours HT offered 75% global histological neuropro-tection with similarly improved regional neuroprotection: thalamus (100%), brainstem (100%), white matter (86%),basal ganglia (76%), cortical gray matter (74%), cerebellum (73%), and hippocampus (72%). Neurology scoresimproved in the 24-hour HT and combined xenon HT groups at 72 hours.Interpretation: Combining xenon with HT is a promising therapy for severely encephalopathic infants, doublingthe neuroprotection offered by HT alone.

ANN NEUROL 2010;68:330–341

Moderate therapeutic hypothermia (HT) to a rectaltemperature of 33 to 35 °C (Trec33-35°C) reduces

brain injury and decreases death/disability with improvedintact neurology following experimental1,2 and clinicalperinatal asphyxia.3–5 However, the frequency of deathand disabilities remains high in these cooled infants (44–55%),3–5 necessitating the development of additional neu-roprotective therapies.

Xenon (Xe), a rare, expensive anesthetic gas with nodocumented adverse effects,6–8 possesses neuroprotectiveproperties by inhibiting N-methyl-D-aspartate (NMDA)receptors9 and other subtypes of glutamate receptors,10

and by reducing apoptosis.11 Because Xe is chemically in-ert with low solubility, its clinical effects can be very rap-idly reversed by stopping gas delivery and allowing it tobe exhaled unchanged. Combining Xe50% and HT32°C inimmature rats offered 71% neuroprotection and long-lasting functional improvement12 compared to 52% pro-tection with HT alone. However, there are limitations tothis unilateral carotid ligation/hypoxia small (12g) neona-tal rat model.13 It produces a pure stroke-like cerebral in-sult without injuring the rest of the body; hence, systemictreatment effects, cardiovascular and biochemical changescannot be assessed. Hence, the neuroprotective effect of

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.22016

Received Nov 3, 2009, and in revised form Feb 17, 2010. Accepted for publication Feb 26, 2010.

Address correspondence to Dr Dingley, Swansea Medical School, University of Wales Swansea, Singleton Park, Swansea, SA2 8PP,United Kingdom.

From 1Department of Clinical Sciences at South Bristol, University of Bristol, Bristol; 2Department of Anaesthetics, Morriston Hospital, Swansea,Wales; and 3Department of Histopathology, Leicester Royal Infirmary, Leicester, United Kingdom.

ORIGINAL ARTICLE

330 © 2010 American Neurological Association

Xe needs to be tested in a large newborn model involvinga global hypoxic-ischemic (HI) insult such as this pigmodel.14 In newborn pigs, cardiovascular physiology,15,16

brain injury,14,17 organ injury,18,19 and neuroprotectiveresponses20,21 following a global HI insult closely resem-ble those seen in human infants after perinatal asphyxia.

The scarcity and high cost ($30/l)22 of Xe are al-leviated by using our closed circle delivery system devel-oped for infants and validated in our pig model, con-suming only 0.10l Xe/kg/h.23 In this study, we reportthe neuroprotective effects in 6 experimental groups sub-jected post-HI insult to NT or 2 different durations ofHT (12 or 24 hours) in combination with 2 breathing gas-eous Xe regimens (0% Xe or 18 hours of 50% Xe):Xe0%NT38.5°C, 18hXe50%NT38.5°C, Xe0%12hHT33.5°C,Xe0%24hHT33.5°C, 18hXe50%12hHT33.5°C, and 18hXe50%

24hHT33.5°C in the global HI encephalopathy (HIE) new-born pig model.

Materials and MethodsConduct of ExperimentThe protocol was conducted under UK Home Office licenseapproved by the University of Bristol Ethical Review Panel. Onehundred thirty-four crossbred Landrace/large white newbornpigs were used; 24 were excluded from final analysis because ofearly death following cardiovascular or respiratory collapse sec-ondary to the severity of the HI insult.

Anesthesia and VentilationAfter unrestrained induction with 2% isoflurane, 65% nitrousoxide (N2O), and 33% oxygen(O2), animals were intubatedwith a 3.0mm cuffed endotracheal tube (Mallinckrodt Medical,Athlone, Ireland) and ventilated (SLE 2000; SLE, Surrey, UK)with 1 to 2% isoflurane, 70% N2O, and 28-29% O2 in theprone position for 60 minutes. Isoflurane was replaced by 0.7%halothane 20 minutes prior to HI insult. After HI insult, inha-lation anesthesia was replaced by intravenous (iv) anesthesia inall animals: a bolus dose of propofol 4mg/kg, followed by main-tenance propofol (4–12mg/kg/h) and remifentanil (20–80�g/kg/h). This iv anesthesia was adjusted according to the heartrate, blood pressure, and responsiveness of the pigs. In the 3 Xetreatment groups, propofol was discontinued once Xe hadreached 50%. Remifentanil was continued during Xe adminis-tration, and the propofol/remifentanil combination resumed af-ter the Xe administration period had ended until extubation. Inthe 3 groups not receiving Xe, propofol/remifentanil iv anesthe-sia was delivered during the entire postinsult period until extu-bation.

Animal PreparationA rectal temperature probe (reusable YSI 400 series, CritiCool,MTRE, Yavne, Israel) was inserted 6cm into the rectum, and askin probe (CritiCool, MTRE) was sited on the ear lobe. Cali-bration was checked for each probe to within �0.1°C over a

temperature range of 20 to 40°C against a certified mercury-in-glass thermometer (BS593; Zeal, London, UK). An ear vein wascannulated, and umbilical arterial and venous catheters were in-serted allowing continuous monitoring of mean arterial bloodpressure (MABP), measurement of blood gases, and infusion ofmaintenance fluids and drugs. An artificial fontanelle (2 � 2cm)was made, as unlike a human neonate, the newborn pig has nopatent fontanelle but open sutures. Probes measuring cortical(0.3cm from the brain surface) and deep gray matter (2.5cmfrom the surface) brain temperatures (Thermes, Physitemp In-struments, Clifton, NJ) were inserted.

Continuous MonitoringDuring NT, rectal temperature (Trec) was maintained at 38.5 �

0.2°C, using the overhead heater in an open Giraffe incubator(Giraffe Omnibed, GE Healthcare, Bucks, UK) (n � 35) or awrap containing circulating water servocontrolled to Trec 38.5°C(CritiCool, MTRE) (n � 75). Oxygen saturation (SpO2),Trec (CritiCool, MTRE), end tidal CO2 (Tidal Wave SP,Respironics Novametrix, Wallingford, CT), and MABP (ICMsoftware, Neurosurgery Unit, University of Cambridge, Cam-bridge, UK) were continuously monitored and recorded. Theventilator settings were altered to maintain the SpO2 between 95and 98% and the end tidal CO2 between 4.5 and 5.5kPa.

Blood gases (analyzed at 37°C) (i-STAT, Abbott Point ofCare, Birmingham, UK; RapidLab 248 pH/blood gas analyzer,Siemens Healthcare Diagnostics, Surrey, UK), blood glucose(Optium, MediSense, Abingdon, UK), and lactate (Lactate Pro,Arkray, Kyoto, Japan) were sampled at preset time points atbaseline and during and after HI insult until 72 hours, as well aswhen clinically indicated.

Electroencephalographic and Amplitude-Integrated ElectroencephalographicMonitoringA single or 2-channel amplitude-integrated electroencephalo-gram (aEEG) (Olympic or Brainz, Natus Medical Incorporated,San Carlos, CA) was continuously recorded with either 2 or 4biparietal (interelectrode distance, 3cm) subdermal needle elec-trodes (019-409700, 0.4mm [27G], Viasys Healthcare, Madi-son, WI).

Fluids and AntibioticsPrior to the HI insult, 3.75% dextrose/0.45% saline, and afterthe HI insult, 5% dextrose/0.45% saline along with propofoland remifentanil were commenced, maintaining the total fluidintake at 12ml/kg/h. Antibiotics (amoxicillin 30mg/kg/12h andgentamicin 5mg/kg/24h) were given for 72 hours.

HI InsultThe cuff of the tracheal tube was inflated and checked for leak-age. Fraction of inspired oxygen (FiO2) was reduced to 5 to 7%to depress the background aEEG activity to below 7�V for 45minutes.14,21 During this 45-minute HI insult, the duration inminutes of low-amplitude background activity �7�V (LAEEG)

Chakkarapani et al: Xenon Enhances Neuroprotection

September, 2010 331

and longest duration of LAEEG14 were measured as markers ofinsult severity.

Post HI InsultAfter resuscitation in air, pigs ventilated with 21 to 30% oxygen,maintaining SpO2 between 95 and 98%, were randomized im-mediately from computer-generated groups and selected fromconsecutive sealed envelopes to 6 treatment groups (Fig 1) or tothe sham group (Fig 2). The sham group underwent similar pro-tocols to experimental animals, except the HI insult to assess theeffect of 18hXe50%, 24hHT33.5°C, and the combination of bothon the uninjured but anesthetized brain.

NTNT was achieved using a radiant warmer (n � 8 pigs) or theCritiCool body wrap (n � 10 pigs). Mean Trec for the radiantwarmer and CritiCool group were 38.7°C and 38.6°C, respec-tively.

HTAfter the HI insult, Trec 33.5°C was achieved within 40 minutesand maintained for 12 or 24 hours using either of 2 methods;fluid-cooled mattress (Tecotherm, Munich, Germany) (n � 18)or CritiCool body wrap (n � 50). After this HT period, re-warming was completed over 10 hours while animals were anes-thetized.

Xenon Delivery and Xe-HT CombinationAfter the HI insult, lungs were ventilated for a runoff period of30 minutes with an air-oxygen mixture to eliminate halothaneprior to commencing the closed circuit xenon delivery period. Axenon/oxygen/nitrogen mixture (Xe 50%, O2 30%, N2 20%) wasthen delivered as previously described23 for a period of 18 hourscombined with either NT throughout or HT for 12 or 24hours. Consequently, rewarming after 12 hours HT was underxenon and iv anesthesia, whereas after 24 hours HT it was un-der iv anesthesia, as the concurrent Xe administration periodwas 18 hours.

Intensive Care Management of All AnimalsHypotension (MABP �40mmHg for �10 minutes)15 wastreated in a stepwise manner after confirming adequate level of

FIGURE 1: Experiment design. Experiment: 98 pigs wererandomized to 6 treatment groups after a 45-minutehypoxic-ischemic insult. Amplitude-integrated electroen-cephalography (aEEG) demonstrates a background ampli-tude of <7�V during the hypoxic-ischemic insult when theinspired oxygen fraction is reduced to 5 to 7%. Black: nor-mothermia (NT) with rectal temperature (Trec) maintainedat 38.5°C. Light gray: 18 hours of 50% Xe along with 30%O2 and 20% N2 delivered via a closed circle. Dark gray:hypothermia (HT) was achieved by whole body coolingwith Trec maintained at 33.5°C. Gray: Rewarming at0.5°C/h. Brains were obtained after terminal perfusion fix-ation. The last bar shows the proportion of brain infarctedand the corresponding neuropathology score. The scale atthe bottom of the figure shows the time line of the ex-periment. Xe0%NT38.5°C (Trec 38.5 � 0.2°C, n � 18),18hXe50%NT38.5°C (18 hours 50% Xe, 30% O2, 20% N2 with72 hours NT, n � 12), Xe0%12hHT33.5°C (12 hours Trec

33.5°C � 10 hours rewarming � 50 hours NT Trec 38.5°C,n � 18), 18hXe50%12hHT33.5°C (18 hours 50% Xe, 30% O2,20% N2 with 12 hours Trec 33.5°C � 10 hours rewarming �50 hours NT Trec 38.5°C, n � 18), Xe0%24hHT33.5°C (24 hoursTrec 33.5°C � 10 hours rewarming � 38 hours NT Trec 38.5°C,n � 17), and 18hXe50%24hHT33.5°C (18 hours 50% Xe, 30%O2, 20% N2 with 24 hours Trec 33.5°C � 10 hours rewarm-ing � 38 hours NT Trec 38.5°C, n � 15). Fio2 � fraction ofinspired oxygen. [Color figure can be viewed in the onlineissue, which is available at www.interscience.wiley.com.]

FIGURE 2: Experiment design for sham experiments.Twelve pigs underwent all the procedures carried out inthe experimental groups except the reduction of inspiredoxygen fraction during the insult period. Amplitude-integrated electroencephalogram (aEEG) shows the ampli-tude >7�V. Black: normothermia (NT) with rectal temper-ature (Trec) maintained at 38.5°C. Light gray: 18 hours of50% Xe with 30% O2 and 20% N2 delivered via a closedcircle. Dark gray: hypothermia (HT) was achieved by wholebody cooling with Trec maintained at 33.5°C. Gray: Re-warming at 0.5°C/h. Brains were obtained after terminalperfusion fixation. The last bar shows the proportion ofbrain infarcted and the corresponding neuropathologyscore. The scale at the bottom of the figure shows thetime line of the experiment. Xe0%NT38.5°C (Trec 38.5 �0.2°C, n � 3), 18hXe50%NT38.5°C (18 hours 50% Xe, 30%O2, 20% N2 with 72 hours NT, n � 3), Xe0%24hHT33.5°C (24hours Trec 33.5°C � 10 hours rewarming � 38 hours NTT

rec38.5°C, n � 3), and 18hXe50%24hHT33.5°C (18 hours

50% Xe, 30% O2, 20% N2 with 24 hours Trec 33.5°C � 10hours rewarming � 38 hours NT Trec 38.5°C, n � 3). Fio2 �fraction of inspired oxygen.

ANNALS of Neurology

332 Volume 68, No. 3

anesthesia with 2 separate 10ml/kg boluses of 0.9% saline fol-lowed by dopamine (5-20�g/kg/min), noradrenaline (20ng–1�g/kg/min), and hydrocortisone (2.5mg/kg/6h or stoppedearly if the blood pressure recovered). Seizures were identifiedon continuous EEG/aEEG recordings. The protocol for treatingboth clinical and electrical seizures lasting �5 minutes was phe-nobarbitone (20mg/kg) � 2, clonazepam (100�g/kg), and lido-caine (2mg/kg over 10 minutes, followed by a continuous infu-sion of 6mg/kg/h for 6 hours, 4mg/kg/h for 12 hours, and 2mg/kg/h for 12 hours).24 Failure to respond to a treatment in 30minutes entailed moving to the next step of this treatment pro-tocol.

Hypoglycemia (blood sugar �54mg/dl [�3mmol/l]) wastreated with a 2.5ml/kg bolus of 10% dextrose, followed by in-creasing the glucose delivery up to 12 mg/kg/min. Persistent hy-perglycemia (blood glucose �180mg/dl) was treated by reducingthe glucose delivery and monitoring the glucose hourly; insulinwas not needed.

Tracheal suction was performed every 12 hours or as clin-ically indicated. Urine was either expressed or tapped by inter-mittent suprapubic aspiration. At the end of the treatment pe-riod in each group, the intracerebral probes were removed, andanimals were weaned from intravenous anesthetics. Pigs were ex-tubated when able to breathe without support, oropharyngealcontinuous positive airway pressure sometimes being used in theinterim. Intramuscular buprenorphine (3�g/kg/8h) was used forpostanesthesia analgesia.

Clinical Neurology AssessmentWe used our previously validated 11-item neurology score (eachitem scored as either 0, very abnormal; 0.5, moderately abnor-mal; 1, mildly abnormal; 2, normal; total score, 0–22), whichwas performed before the start of the experiment and at 72hours after HI insult.14,19 Anesthesia, sedation, and 72-hoursurvival preclude any cognitive assessment within the neurolog-ical examination.

Neuropathology AssessmentAfter 3 days survival, pigs were reintubated and, under deepisoflurane anesthesia, underwent terminal perfusion fixation ofthe brain with 0.9% saline through the internal carotid arteriesfollowed by 4% phosphate-buffered formaldehyde and a full au-topsy. Hematoxylin & eosin stained sections obtained from ev-ery 5mm of the right uninstrumented hemisphere were assessedin 7 brain regions (cortex gray matter, white matter, basal gan-glia, thalamus, hippocampus, cerebellum, and brainstem) by aperinatal pathologist blinded to the randomization and clinicaldetails using an established 9-point neuropathology score at in-tervals of 0.5 from 0.0 to 4.0.14 Global neuropathology was cal-culated by averaging the regional neuropathology scores. Theneuropathology score corresponding to the proportion of brainarea damaged14 is shown in Figure 1.

Statistical AnalysisContinuous and skewed variables were summarized as mean(standard deviation [SD]) and median (95% confidence interval

[CI]), respectively (Minitab Software, Coventry, UK). Differ-ences in the normally distributed estimates were assessed with1-way analysis of variance, and skewed estimates were assessedwith Kruskal-Wallis or median test. Spearman correlation wasused to assess the association between the clinical neurologyscore and global neuropathology. Interaction between xenon andHT was assessed using univariate full factorial linear models.HT was graded as NT (0), 12 hours HT (1), or 24 hours HT(2). We investigated between-subject effects for the dependentvariables average global neuropathology and regional neuropa-thology. The design was comprised of intercept � HT33.5°C

graded (0/1/2) � 18hXe50% (0/1 � no/yes) � HT33.5°C

graded � 18hXe50%, where the term HT33.5°C graded �

18hXe50% represents the question, does the combined effect ofthese 2 treatments differ from the sum of the effects of eachwhen given alone? The independent effect of treatment with 18hours Xe50%, 12 hours HT, and 24 hours HT on neuropathol-ogy was assessed using multiple linear regression, and performedagain with insult duration (duration of low amplitude [�7�V]aEEG in minutes), pH (�7 or �7), and lactate at the end ofHI insult as additional covariates to control for the severity ofthe insult (SPSS version 15, SPSS Inc., Chicago, IL).

The independent effects of different cooling equipment(Tecotherm or CritiCool), duration of inhalational anesthesia,and propofol dose were assessed using multiple linear regression,with global neuropathology as a dependent variable. A probabil-ity value of �0.05 (2-sided) was considered statistically signifi-cant.

ResultsThere was no difference in the weight, age, sex distribu-tion, or insult severity between the groups (Table 1).

Average Brain Pathology18hXe50% and HT33.5°C (12 hours HT/24 hours HT)had a significant protective effect on global and regionalneuropathology (Table 2). There was no interaction be-tween 18hXe50% and the 2 different durations of HT, asshown by the insignificant p values of “Xe*HT graded,”indicating an additive neuroprotective effect between18hXe50% and HT, rather than a synergistic effect (seeTable 2).

18hXe50%, 12hHT33.5°C, and 24hHT33.5°C therapysignificantly reduced the global and all the regional neu-ropathology scores (Table 3). When the significantindependent effects of 18hXe50%, 12hHT33.5°C, and24hHT33.5°C treatments were assessed with variables rep-resenting insult severity (LAEEG, arterial lactate, and pH[� 7 or �7] at the end of HI insult) as covariates, theprediction improved only slightly. For every 1mmol/l in-crease in lactate, there was an independent significant in-crease in the global, cortical gray matter, white matter,hippocampus, and cerebellum neuropathology scores by0.146, 0.174, 0.137, 0.143, and 0.132, respectively.

Chakkarapani et al: Xenon Enhances Neuroprotection

September, 2010 333

The micrographs of cortex and hippocampus in Fig-ure 3 show injury grades of 0, 1, 2, 3, and 4 from animalsof the 18hXe50%24hHT33.5°C, Xe0%24hHT33.5°C,18hXe50%NT38.5°C, Xe0%12hHT33.5°C, and Xe0%NT38.5°C

groups, respectively. The 18hXe50%24hHT33.5°C group hadthe greatest reduction in global brain injury neuropathol-ogy score: neuropathology score 0.53 vs Xe0%24hHT33.5°C,1.25; 18hXe50%12hHT33.5°C, 1.45; 18hXe50%NT38.5°C,1.71; Xe0%12hHT33.5°C, 2.02; and Xe0%NT38.5°C, 2.39(Fig 4). The observed and estimated global neuropathologyscore correlated well, as shown by the estimated score lyingwithin the observed score error bars (see Fig 4). Whereas24hHT provided 48% neuroprotection, combining 18hXewith 24hHT offered 75% neuroprotection, and this effectwas additive. The 18hXe50%24hHT33.5°C group was able toimprove outcome in severely brain-injured pigs (Fig 5).Three outliers in the 18hXe50%24hHT33.5°C and 2 of the 4outliers in the Xe0%24hHT33.5°C and 18hXe50%

12hHT33.5°C groups had status epilepticus (odds ratio,10.8; 95% CI, 1.1–121.0) as well as long duration of low-amplitude EEG during the HI insult (for every minute in-crease in the longest duration of LAEEG, the odds of hav-ing a global neuropathology score �2 increased by 1.14;95% CI, 1.02–1.26). There was no significant difference in

the incidence of seizures between any of the groups (me-dian incidence, 17%; interquartile range, 11.8–21.8%).

Regional Brain PathologyThe 18hXe

50%24hHT33.5°C group had significant neuro-

protection in all brain regions (Fig 6). When the esti-mated regional neuropathology score of 18hXe50%

24hHT33.5°C (see Table 3) was normalized to NT, theprotection was 100% in the thalamus (no injury), 100%in brainstem, 86% in white matter, 76% in basal ganglia,74% in cortical gray matter, 73% in cerebellum, and72% in hippocampus; this correlated with the observedregional neuropathology scores (see Fig 6).

Clinical Neurology ScoreThe 18hXe50%12hHT33.5°C, Xe0%24hHT33.5°C, and18hXe50%24hHT33.5°C groups showed significantly betterneurology scores at 72 hours compared to the Xe0%

NT38.5°C, 18hXe50%NT38.5°C, and Xe0%12hHT33.5°C

groups (p � 0.0002) (Fig 7). Neurology scores at 72 hourscorrelated more significantly with the global neuropathologyin the NT groups (Xe0%NT38.5°C, r � �0.81;18hXe50%NT38.5°C, r � �0.83) than in the HT groups(Xe

0%12hHT33.5°C, r � �0.06; Xe0%24hHT33.5°C, r �

TABLE 1: Characteristics of Pigs and Hypoxic-Ischemic Insult in the 6 Treatment Groups

GroupVariable

Treatment p

Xe0%

NT38.5°C,n � 18

18hXe50%

NT38.5°C,n � 12

Xe0%

12hHT33.5°C,n � 18

18hXe50%

12hHT33.5°C,n � 18

Xe0%

24hHT33.5°C,n � 17

18hXe50%

24hHT33.5°C,n � 15

Weight, g, mean(95% CI)

1,550(1,419-1,681)

1,673(1,510-1,835)

1,720(1,626-1,815)

1,643(1,523-1,764)

1,702(1,561-1,843)

1,738(1,569-1,906)

0.32

Age, h, median(95% CI)

18(13.5-22.5)

17.1(10.5-20.5)

16.5(12-20.5)

20.4(16.5-25.15)

19.5(13.5-24)

14(10-22.8)

0.49

Female 50% 66% 50% 61% 47% 34% 0.24

Duration ofLAEEG duringinsult, min,mean (95% CI)

37.14(33.73-40.54)

32.77(25.78-39.76)

33.85(30.88-36.82)

31.73(26.81-36.66)

30.62(26.01-35.23)

33.45(29.62-37.28)

0.258

HI end: bloodlactate, mmol/l,median(95% CI)

16.1(14.9-17.2)

15.5(13.1-18.1)

17.4(16.3-18.4)

17.7(16.5-18.6)

16.2(14.9-17.7)

15.4(14.2-16.1)

0.1

HI end: arterialpH, mean(95% CI)

7.03(6.96-7.11)

7.13(7.03-7.23)

7.03(6.95-7.10)

7.09(7.04-7.15)

7.05(6.98-7.12)

7.10(7.04-7.16)

0.30

Duration of low-amplitude encephalography, arterial lactate, and pH at the end of insult indicate the severity of hypoxic-ischemic insult.Xe � xenon; NT � normothermia; HT � hypothermia; CI � confidence interval; LAEEG � low-amplitudeelectroencephalography; HI � hypoxic-ischemic insult.

ANNALS of Neurology

334 Volume 68, No. 3

TABLE 2: Treatment Effects on Global and Regional Neuropathology

Source Dependent Variable Type III Sum of Squares df F p

Corrected model Global neuropathology 35.331 5 4.404 0.001

Gray matter 30.447 5 3.466 0.006

White matter 40.872 5 4.524 0.001

Basal ganglia 38.254 5 3.191 0.011

Thalamus 36.167 5 5.143 0.0001

Hippocampus 35.144 5 3.310 0.009

Cerebellum 37.704 5 3.385 0.007

Brainstem 5.549 5 3.396 0.007

Intercept Global neuropathology 220.294 1 137.286 0.0001

Gray matter 277.885 1 158.148 0.0001

White matter 217.942 1 120.604 0.0001

Basal ganglia 186.781 1 77.905 0.0001

Thalamus 53.631 1 38.132 0.0001

Hippocampus 335.254 1 157.889 0.0001

Cerebellum 336.887 1 151.211 0.0001

Brainstem 4.078 1 12.479 0.001

18hXe50% Global neuropathology 10.811 1 6.737 0.011

Gray matter 12.729 1 7.244 0.008

White matter 14.391 1 7.964 0.006

Basal ganglia 8.195 1 3.418 0.068

Thalamus 8.514 1 6.053 0.016

Hippocampus 10.269 1 4.836 0.030

Cerebellum 11.161 1 5.010 0.028

Brainstem 1.878 1 5.749 0.019

HT graded (NT/12hHT/24hHT) Global neuropathology 19.321 2 6.020 0.003

Gray matter 13.669 2 3.890 0.024

White matter 19.542 2 5.407 0.006

Basal ganglia 22.046 2 4.598 0.012

Thalamus 17.036 2 6.057 0.003

Hippocampus 20.973 2 4.939 0.009

Cerebellum 23.543 2 5.283 0.007

Brainstem 2.264 2 3.464 0.036

Xe*HT graded Global neuropathology 1.999 2 0.623 0.539

Gray matter 1.520 2 0.433 0.650

White matter 2.665 2 0.737 0.481

Basal ganglia 5.517 2 1.150 0.321

Thalamus 6.306 2 2.242 0.112

Hippocampus 0.971 2 0.229 0.796

Cerebellum 0.504 2 0.113 0.893

Brainstem 1.070 2 1.637 0.200

Results from univariate full factorial linear model investigating between-subject effects for 2 dependent variables: global andregional neuropathology.df � degrees of freedom; Xe � xenon; HT � hypothermia; NT � normothermia.

Chakkarapani et al: Xenon Enhances Neuroprotection

September, 2010 335

�0.45; 18hXe50%12hHT33.5°C, r � �0.53; 18hXe50%

24hHT33.5°C, r � �0.26).

Temperature ManagementIn the Xe0%12hHT33.5°C group, the target temperature of33.5°C was reached by mean (SD) 36 (2.41) minutes;mean (95% CI) Trec was 33.4 (33.19–33.56) °C for 12hours, and the rewarming period was mean (95% CI)9.07 (7.82–10.32) hours (Fig 8). Mean (95% CI) super-ficial and deep brain temperatures during the 12-hourcooling period were 31.9 (31.7–32.16) °C and 33.4(33.24–33.51) °C, respectively. In the Xe0%24hHT33.5°C

group, mean (95% CI) Trec was 33.5 (33.41–33.49) °Cfor 24 hours, and the rewarming period was 10.06 (8.6–11.6) hours. Mean (95% CI) superficial and deep braintemperatures during the 24hHT period were 31.8(31.62–31.97) °C and 33.5 (33.37–33.52) °C, respec-

tively (see Fig 8). The cooling equipment used (Teco-therm or CritiCool) did not influence the neuropathology(p � 0.73).

Anesthesia, Xe Concentration,and ConsumptionThere was no difference in the duration of preinsult inha-lation anesthesia with nitrous oxide and isoflurane, postin-sult propofol, remifentanil, Xe concentration, and Xe con-sumption between the relevant treatment groups (data notshown). Median (95% CI) inspired Xe concentrations dur-ing 18 hours of Xe therapy were: Xe50%NT38.5°C, 53.4%(52.3–54.6%), Xe50%12hHT33.5°C, 52.1% (51.4–52.8%),and Xe50%24hHT33.5°C, 52.3% (51.5–53.1%). Median(95% CI) inspired oxygen fraction during 18 hours of Xetherapy for Xe50%NT38.5°C was 30.2% (29–31.5%), forXe50%12hHT33.5°C was 34.0% (33.0 –35.0%), and for

TABLE 3: Multiple Linear Regression Showing the Independent Effect of Xenon and Hypothermia

Model Coefficient SE p Lower 95% CI Upper 95% CI R2

Global neuropathology constant 2.442 0.234 0.0001 1.977 2.906 0.18

HT (0/12h/24h) �0.581 0.161 0.0001 �0.901 �0.262

Xe18h (yes/no) �0.680 0.255 0.009 �1.186 �0.174

Cortical gray matter constant 2.561 0.244 0.0001 2.076 3.046 0.15

HT (0/12h/24h) �0.485 0.168 0.005 �0.819 �0.152

Xe18h (yes/no) �0.732 0.266 0.007 �1.260 �0.203

Cortical white matter constant 2.499 0.248 0.001 2.006 2.992 0.18

HT (0/12h/24h) �0.596 0.171 0.001 �0.935 �0.256

Xe18h (yes/no) �0.789 0.271 0.004 �1.326 �0.251

Basal ganglia constant 2.294 0.287 0.0001 1.723 2.864 0.13

HT (0/12h/24h) �0.597 0.198 0.003 �0.990 �0.205

Xe18h (yes/no) �0.625 0.313 0.04 �1.247 �0.004

Thalamus constant 1.627 0.223 0.0001 1.185 2.069 0.18

HT (0/12h/24h) �0.572 0.153 0.001 �0.876 �0.267

Xe18h (yes/no) �0.597 0.243 0.016 �1.078 �0.115

Hippocampus constant 2.810 0.268 0.0001 2.279 3.342 0.15

HT (0/12h/24h) �0.605 0.184 0.001 �0.971 �0.239

Xe18h (yes/no) �0.656 0.292 0.027 �1.236 �0.077

Cerebellum constant 2.853 0.274 0.0001 2.308 3.398 0.15

HT (0/12h/24h) �0.629 0.189 0.001 �1.004 �0.254

Xe18h (yes/no) �0.674 0.299 0.026 �1.267 �0.08

Brainstem constant 0.504 0.109 0.0001 0.287 0.720 0.09

HT (0/12h/24h) �0.148 0.075 0.053 �0.298 0.002

Xe18h (yes/no) �0.276 0.120 0.024 �0.514 �0.037

Dependent variables: global and regional neuropathology.SE � standard error; CI � confidence interval; HT � hypothermia; Xe � xenon.

ANNALS of Neurology

336 Volume 68, No. 3

FIGURE 3: Micrographs of cortex and hippocampus showing injury grades from 0 to 4. Grade 0, 1, 2, 3, and 4 are from animals of18hXe50%24hHT33.5°C, Xe0%24hHT33.5°C, 18hXe50%NT38.5°C, Xe0%12hHT33.5°C, and Xe0%NT38.5°C, respectively. Black arrows indicate de-generating neurons and neuronal loss. Blue arrows indicate normal neurons in the cortex and hippocampus. Grade 0 shows normalcortex and hippocampus. Grade 1 shows scattered degenerating neurons in otherwise normal cortex or hippocampus. Grade 2 showsmore degenerating neurons than grade 1, some of which are in small groups. Grade 3 in cortex shows neuronal degeneration andneuronal loss at the base of the sulcus (ulegyria). Grade 3 in hippocampus shows degeneration of most of the neurons in the section.Grade 4 in cortex shows extensive area of neuronal degeneration and neuronal loss with a small area of normal cortex. Grade 4 inhippocampus shows degeneration of all neurons. HT � hypothermia; NT � normothermia.

Chakkarapani et al: Xenon Enhances Neuroprotection

September, 2010 337

Xe50%24hHT33.5°C was 31.5% (30.4 –32.7%). Overallonly 4.5l Xe was used per animal during each 18-hourtreatment.

DiscussionIn our global HIE pig model, the combination of Xe50%

and 24-hour HT to 33.5°C offered 75% global neuropro-

tection. The neuroprotective effects of Xe-HT were addi-tive when administered together. With this combinationtherapy, each brain region had significantly less histologicalinjury than with either treatment alone. We also documentbetter neuroprotection with increasing duration of HT, 24-hour HT being more protective than 12-hour HT.

In this study, we examined whether the additiveneuroprotection of the Xe-HT combination that we pre-viously demonstrated in the Vannucci13 rat model12

could be reproduced in our larger, more clinically rele-vant, newborn global HI pig model. In the univariate fullfactorial linear model, there was no interaction between18hXe50% and HT33.5°C, indicating that the treatment ef-

FIGURE 4: Average global neuropathology in the treatmentgroups and sham animals. The figure shows the median (95%confidence interval) average global neuropathology score (lightgray); estimated average global neuropathology based on thelinear regression model (gray) 2.442–0.581 * hypothermia [HT]�0.680 * Xe; HT is coded 0 for normothermia [NT], 1 for 12-hour HT, 2 for 24-hour HT. Xe is coded 0 for no Xe therapy and1 for 18-hour Xe therapy. Estimated average global neuropa-thology adjusted for insult severity (black) �0.056–0.579 * HT�0.607 * Xe � 0.146 * end insult lactate. Insult severity is rep-resented by duration of amplitude-integrated electroencephalo-gram <7�V during the 45-minute hypoxic-ischemic insult, lac-tate, and arterial pH at the end of the hypoxic-ischemic insult.The following abbreviations in the figure correspond to thetreatment groups as indicated. NT � Xe0%NT38.5°C; 24hHT �Xe0%24hHT33.5°C; XeNT � 18hXe50%NT38.5°C; Xe24hHT �18hXe50%24hHT33.5°C; 12hHT � Xe0%12hHT33.5°C; Xe12hHT �18hXe50%12hHT33.5°C.

FIGURE 5: Distribution of individual and mean averageglobal neuropathology score (0 � no injury, 4 � maximuminjury) of animals in all treatment groups indicated with indi-vidual symbols. Circles � Xe0%NT38.5°C; squares � 18hXe50%

NT38.5°C; open triangles � Xe0%12hHT33.5°C; open dia-monds � 18hXe50%12hHT33.5°C; solid triangles � Xe0%

24hHT33.5°C; solid diamonds � 18hXe50%24hHT33.5°C. NT �normothermia; HT � hypothermia.

FIGURE 6: Distribution of regional neuropathology betweengroups. Bars with error bars indicate the median (95% confi-dence interval) of observed data and the neighboring bars with-out error bars are the estimated data based on the regressionmodel. Black: Xe0%NT38.5°C; very light gray: Xe0%12hHT33.5°C;gray: 18hXe50%NT38.5°C; dark gray: 18hXe50%12hHT33.5°C; lightgray: Xe0%24hHT33.5°C and; white: 18hXe50%24hHT33.5°C. NT �normothermia; HT � hypothermia; GM � cortical gray matter;WM � white matter; BG � basal ganglia; THAL � thalamus;HIPPO � hippocampus; CBM � cerebellum; BS � brainstem.

FIGURE 7: Median (95% confidence interval) clinical neurologyscore at 72 hours in the treatment groups. Clinical neurologyscore is an 11-item score (0 � abnormal; 2 � normal; range,0–22), and higher scores represent better clinical neurology. Thefollowing abbreviations in the figure correspond to the treat-ment groups as indicated. NT � Xe0%NT38.5°C; XeNT �18hXe50%NT38.5°C; 12hHT � Xe0%12hHT33.5°C; Xe12hHT �18hXe

50%12hHT33.5°C 5; 24hHT � Xe0%24hHT33.5°C; Xe24hHT �

18hXe50%24hHT33.5°C.. NT � normothermia; HT � hypothermia.

ANNALS of Neurology

338 Volume 68, No. 3

fect of the Xe-HT combination was not different fromthe sum of the treatment effects of Xe and HT alone,demonstrating an additive effect. Had there been a syner-gistic effect as suggested,11 where a low dose, 20% of Xe(not neuroprotective on its own) was combined with verymild nonprotective HT, the combination treatment wasneuroprotective; the combined Xe-HT effect would belarger than the sum, and the interaction between Xe andHT would have been significant. On applying this addi-tive effect, the regression model offered 48%, 28%, and75% neuroprotection with 24hHT33.5°c, 18hXe50%, and18hXe50%24hHT33.5°C, respectively. Although Xe11 andHT25,26 may share neuroprotective mechanisms, it islikely that they differ in their main mechanism of ac-tion.9,20,27–32 The timing at which they influence theirrespective therapeutic pathways may add to the effective-

ness. Our study suggests there is room for further neuro-protection after the therapeutic effect of HT has beenfully maximized as regards depth and duration. There is arole for prudent use of an expensive and scarce gas like Xeto enhance neuroprotection. We have now confirmed anadditive Xe-HT effect both in the rat carotid ligation/hypoxia model,12 and a newborn pig global HIE model,and this effect is very likely to translate to humans. Ourprevious work on the neuroprotective effect of HT trans-lated from rat2,33 to pig21,34 to human.3,5

Xenon50% combined with 24hHT33.5°C offered re-gional neuroprotection in all brain regions, with a trendof more protection in the basal ganglia and thalamus.Postasphyxic striatal injury in newborn pigs mediatedthrough NMDA receptor activation is reduced by HT,32

and because Xe is an NMDA inhibitor, this might en-hance striatal neuroprotection. Apoptosis mediates post-asphyxic thalamic injury in rats,35 and is likely to do so inpigs. NMDA receptor activation results in necrosis, and acontinuum is described from apoptosis to necrosis.36

Xe50%11 and HT33.5°C

25 may reduce necrosis and apopto-sis through several mechanisms, yielding additive neuro-protection.

Twelve hours of whole body HT ameliorated sec-ondary energy failure and reduced apoptosis in a carotidocclusion pig HI model,26,37 and improved survival andneurological outcome38 in humans after cardiac arrest. Inour global HIE piglet model, 6-hour HT protectedagainst mild insults,39 whereas 24-hour HT significantlyprotected against both mild and more severe insults.21

There was a significant reduction in global neuropathol-ogy in the 24-hour HT group compared to the 12-hourHT group, showing the duration-dependent neuroprotec-tive effect of HT.

A core temperature of 33.5°C is a 5°C reductionfrom pig NT, and in our experience this is the deepestHT that can be maintained for 24 hours in pigs withoutsystemic adverse effects such as hypotension and signs ofpoor perfusion (metabolic acidosis).

All our treatment groups had a similar duration ofinhalation anesthesia with isoflurane and N2O. Nitrousoxide and isoflurane are shown to induce neuroapoptosisin the developing rat brain.40,41 However, in our model24 hours of inhalation anesthesia (isoflurane) followed byiv anesthesia (fentanyl and propofol) did not increase neu-ronal apoptosis42 during NT or HT.

We could only examine 1 duration of xenon deliv-ery in a 6-group design and chose 18-hour Xe, which ishalfway between 12 and 24 hours of HT. We have pre-viously administered Xe50% successfully for this period us-ing the prototype of the Xe delivery system.23 Previous in

FIGURE 8: Mean (standard error of the mean) of rectal, super-ficial, and deep brain temperature. Xe0%NT38.5°C (black circles:deep brain temperature; blue circles: superficial brain tempera-ture; red circles: rectal temperature); 18hXe50%NT38.5°C (blacksquares: deep brain temperature; blue squares: superficial braintemperature; red squares: rectal temperature); Xe0%12hHT33.5�C

(open black triangles: deep brain temperature; open blue trian-gles: superficial brain temperature; open red triangles: rectaltemperature), 18hXe50%12hHT33.5°C (solid black triangles: deepbrain temperature; solid blue triangles: superficial brain temper-ature; solid red triangles: rectal temperature), Xe0%24hHT33.5°C

(open black diamonds: deep brain temperature; open blue dia-monds: superficial brain temperature; open red diamonds: rectaltemperature), and 18hXe50%24hHT33.5°C (solid black diamonds:deep brain temperature; solid blue diamonds: superficial braintemperature; solid red diamonds: rectal temperature). Duringhypoxic-ischemic insult, the inspired oxygen fraction was re-duced to 5 to 7% until the amplitude-integrated electroenceph-alogram background fell below 7�V, and the rectal temperaturewas maintained at normothermia. Rewarming (gray shading) wasaccomplished at a rate of 0.5°C/h over 10 hours. Superficialbrain temperature was measured at 0.3cm from the surface ofthe brain corresponding to the cortical gray matter, and deepbrain temperature was measured at 2.5cm from the cortex cor-responding to deep gray matter. NT � normothermia; HT �hypothermia.

Chakkarapani et al: Xenon Enhances Neuroprotection

September, 2010 339

vitro and most in vivo Xe work showing significant neu-roprotection has used 70 to 80% Xe.11,43 Such high Xeconcentrations do not offer room to adjust the oxygenconcentration in response to patient need, and may leadto hypoxia if the Xe concentration exceeds 80% by hu-man error. Indeed, asphyxiated infants often need �30%FiO2. In our rodent work, Xe70% produced sedation andrespiratory depression with CO2 retention.44 We thereforechose to use Xe50% in these studies, which was neuropro-tective12,45 but still allowed normal carbon dioxide pres-sure and oxygen pressure both at NT and HT.

Xenon is a rare and expensive gas needing specialdelivery equipment. Our Xe delivery system administered50% Xe in all the groups while maintaining O2 in theclosed circle at �35%, to minimize hyperoxia-or hypoxia-induced secondary brain injury.46 Xe consumption de-creased to 0.1l/h ($3/h at $30/l) as we gained experience.

There are some limitations in our study. The treat-ments were started immediately after HI insult with a rel-atively short survival period. In the clinical situation, thetime of onset of HI insult is often unknown, and the treat-ment started after a delay. In our study, commencing Xe30 minutes after HI insult, when the pigs had achieved Trec

of 33.5°C, offered significant additive neuroprotection. Thetherapeutic window for Xe’s neuroprotective effect in clin-ical settings is unknown. When as short a period as 1 hourof Xe was added after 2 hours of HT, there was a signifi-cant long-term (11 weeks) neuroprotective and functionalimprovement in the neonatal rat model.47

Although 3 days survival in our model is relativelyshort, our previous study in this model investigating 24hours of selective head cooling21 yielding significant neu-roprotection was replicated in newborn asphyxiated in-fants with significant reduction in death or disability at18 months of age.3–5 We have recent evidence that Xeand HT extend the therapeutic window of the otheragent when combined as shown by equivalent neuropro-tection offered by Xe given in the first of 3 hours of HTand Xe given during the third of 3 hours of HT after HIinsult in p7 rats.47 Although 5-hour HT commenced af-ter a 5-hour delay failed to yield significant neuroprotec-tion, starting the combination Xe-HT treatment with a5-hour delay offered significant neuroprotection (unpub-lished data). This provides clear evidence of Xe’s ability toextend the therapeutic window; however, the precisemechanism is unclear. The principle of staggered and de-layed combination treatment was elegantly shown by Di-etrich et al48 in an adult rodent transient global forebrainischemia model. A short (3 hours) period of immediateHT to 30°C, that in itself was not neuroprotective,showed hippocampal neuroprotection after 2 months

when immediate HT was combined with ultradelayed (3days) pharmacological treatment with an NMDA antago-nist (MK 801). Anesthetics that enhance neuroprotectionexperimentally49 have not been tested for this indicationclinically in newborn infants. Although the experimentalevidence for the neuroprotective effect of Xe-HT is increas-ing, it needs confirmation in clinical trials. The therapeuticwindow and the optimal effective treatment duration of Xethat can be combined with 72-hour therapeutic HT afterperinatal asphyxia are at present unknown.

In conclusion, 18 hours of Xe50% and 24 hours ofwhole body cooling to 33.5°C offered significant reduc-tions of 28% and 48% in global neuropathology, respec-tively. The combination of both treatments is additive,yielding 75% reduction in global neuropathology in aglobal HI newborn pig model that closely resembles hu-man perinatal asphyxia.

AcknowledgmentsThis study was supported by Children’s Medical ResearchCharity (05 BTL 01 to M.T.), Sport Aiding Medical Re-search For Kids (SPARK), United Kingdom.

SLE, CritiCool, and Datex-Ohmeda donated equip-ment for the study.

Potential Conflicts of InterestNothing to report.

References1. Colbourne F, Sutherland G, Corbett D. Postischemic hypother-

mia. A critical appraisal with implications for clinical treatment.Mol Neurobiol 1997;14:171–201.

2. Bona E, Hagberg H, Loberg EM, et al. Protective effects of mod-erate hypothermia after neonatal hypoxia-ischemia: short-andlong-term outcome. Pediatr Res 1998;43:738–745.

3. Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head coolingwith mild systemic hypothermia after neonatal encephalopathy: mul-ticentre randomised trial. Lancet 2005;365:663–670.

4. Shankaran S, Laptook AR, Ehrenkranz RA, et al. Whole-body hy-pothermia for neonates with hypoxic-ischemic encephalopathy.N Engl J Med 2005;353:1574–1584.

5. Azzopardi DV, Strohm B, Edwards AD, et al. Moderate hypother-mia to treat perinatal asphyxial encephalopathy. N Engl J Med2009;361:1349–1358.

6. Marx T, Froeba G, Wagner D, et al. Effects on haemodynamicsand catecholamine release of xenon anaesthesia compared withtotal i.v. anaesthesia in the pig. Br J Anaesth 1997;78:326–327.

7. Dingley J, King R, Hughes L, et al. Exploration of xenon as apotential cardiostable sedative: a comparison with propofol aftercardiac surgery. Anaesthesia 2001;56:829–835.

8. Rossaint R, Reyle-Hahn M, Schulte Am Esch J, et al. Multicenterrandomized comparison of the efficacy and safety of xenon andisoflurane in patients undergoing elective surgery. Anesthesiol-ogy 2003;98:6–13.

9. Franks NP, Dickinson R, de Sousa SL, et al. How does xenonproduce anaesthesia? Nature 1998;396:324.

ANNALS of Neurology

340 Volume 68, No. 3

10. Dinse A, Fohr KJ, Georgieff M, et al. Xenon reduces glutamate-,AMPA-, and kainate-induced membrane currents in cortical neu-rones. Br J Anaesth 2005;94:479–485.

11. Ma D, Hossain M, Chow A, et al. Xenon and hypothermia com-bine to provide neuroprotection from neonatal asphyxia. AnnNeurol 2005;58:182–193.

12. Hobbs C, Thoresen M, Tucker A, et al. Xenon and hypothermiacombine additively, offering long-term functional and histopatho-logic neuroprotection after neonatal hypoxia/ischemia. Stroke2008;39:1307–1313.

13. Rice JE III, Vannucci RC, Brierley JB. The influence of immaturity onhypoxic-ischemic brain damage in the rat. Ann Neurol 1981;9:131–141.

14. Thoresen M, Haaland K, Loberg EM, et al. A piglet survival modelof posthypoxic encephalopathy. Pediatr Res 1996;40:738–748.

15. Haaland K, Karlsson B, Skovlund E, et al. Postnatal developmentof the cerebral blood flow velocity response to changes in CO2and mean arterial blood pressure in the piglet. Acta Paediatr1995;84:1414–1420.

16. Fugelseth D, Satas S, Runde M, et al. Cardiac function and mor-phology studied by two-dimensional Doppler echocardiographyin unsedated newborn pigs. Exp Physiol 1999;84:69–78.

17. Thoresen M, Hallstrom A, Whitelaw A, et al. Lactate and pyru-vate changes in the cerebral gray and white matter during post-hypoxic seizures in newborn pigs. Pediatr Res 1998;44:746–754.

18. Satas S, Loberg EM, Porter H, et al. Effect of global hypoxia-ischaemia followed by 24 h of mild hypothermia on organ pa-thology and biochemistry in a newborn pig survival model. BiolNeonate 2003;83:146–156.

19. Karlsson M, Tooley JR, Satas S, et al. Delayed hypothermia asselective head cooling or whole body cooling does not protectbrain or body in newborn piglets. Pediatr Res 2008;64:74–80.

20. Thoresen M, Satas S, Puka-Sundvall M, et al. Post-hypoxic hypo-thermia reduces cerebrocortical release of NO and excitotoxins.Neuroreport 1997;8:3359–3362.

21. Tooley JR, Satas S, Porter H, et al. Head cooling with mildsystemic hypothermia in anesthetized piglets is neuroprotective.Ann Neurol 2003;53:65–72.

22. Rawat S, Dingley J. Closed-circuit xenon delivery using a stan-dard anesthesia workstation. Anesth Analg 2010;110:101–109.

23. Chakkarapani E, Thoresen M, Hobbs CE, et al. A closed-circuit neo-natal xenon delivery system: a technical and practical neuroprotectionfeasibility study in newborn pigs. Anesth Analg 2009;109:451–460.

24. Malingre MM, Van Rooij LG, Rademaker CM, et al. Developmentof an optimal lidocaine infusion strategy for neonatal seizures.Eur J Pediatr 2006;165:598–604.

25. Wang LS, Yu LJ, Shao XM. Mild hypothermia attenuates neuro-nal apoptosis after cerebral hypoxia-ischemia in neonatal rats [inChinese]. Zhongguo Dang Dai Er Ke Za Zhi 2007;9:37–41.

26. Edwards AD, Yue X, Squier MV, et al. Specific inhibition of apo-ptosis after cerebral hypoxia-ischaemia by moderate post-insult hy-pothermia. Biochem Biophys Res Commun 1995;217:1193–1199.

27. Natale G, Cattano D, Abramo A, et al. Morphological evidencethat xenon neuroprotects against N-methyl-DL-aspartic acid-induced damage in the rat arcuate nucleus: a time-dependentstudy. Ann N Y Acad Sci 2006;1074:650–658.

28. Petzelt C, Blom P, Schmehl W, et al. Prevention of neurotoxicityin hypoxic cortical neurons by the noble gas xenon. Life Sci2003;72:1909–1918.

29. Ma D, Hossain M, Pettet GK, et al. Xenon preconditioning re-duces brain damage from neonatal asphyxia in rats. J CerebBlood Flow Metab 2006;26:199–208.

30. Gruss M, Bushell TJ, Bright DP, et al. Two-pore-domain K�channels are a novel target for the anesthetic gases xenon, ni-trous oxide, and cyclopropane. Mol Pharmacol 2004;65:443–452.

31. Erecinska M, Thoresen M, Silver IA. Effects of hypothermia onenergy metabolism in mammalian central nervous system.J Cereb Blood Flow Metab 2003;23:513–530.

32. Mueller-Burke D, Koehler RC, Martin LJ. Rapid NMDA receptorphosphorylation and oxidative stress precede striatal neurode-generation after hypoxic ischemia in newborn piglets and areattenuated with hypothermia. Int J Dev Neurosci 2008;26:67–76.

33. Thoresen M, Bagenholm R, Loberg EM, et al. Posthypoxic cool-ing of neonatal rats provides protection against brain injury. ArchDis Child 1996;74:F3–F9.

34. Haaland K, Loberg EM, Steen PA, et al. The effect of mild post-hypoxic hypothermia on organ pathology in a piglet survivalmodel of global hypoxia. Prenat Neonatal Med 1997;2:329–337.

35. Northington FJ, Ferriero DM, Flock DL, Martin LJ. Delayed neu-rodegeneration in neonatal rat thalamus after hypoxia-ischemia isapoptosis. J Neurosci 2001;21:1931–1938.

36. Portera-Cailliau C, Price DL, Martin LJ. Non-NMDA and NMDAreceptor-mediated excitotoxic neuronal deaths in adult brain aremorphologically distinct: further evidence for an apoptosis-necrosis continuum. J Comp Neurol 1997;378:88–104.

37. Thoresen M, Penrice J, Lorek A, et al. Mild hypothermia after se-vere transient hypoxia-ischemia ameliorates delayed cerebral energyfailure in the newborn piglet. Pediatr Res 1995;37:667–670.

38. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatosesurvivors of out-of-hospital cardiac arrest with induced hypother-mia. N Engl J Med 2002;346:557–563.

39. Haaland K, Loberg EM, Steen PA, Thoresen M. Posthypoxic hy-pothermia in newborn piglets. Pediatr Res 1997;41:505–512.

40. Jevtovic-Todorovic V, Hartman RE, Izumi Y, et al. Early exposureto common anesthetic agents causes widespread neurodegen-eration in the developing rat brain and persistent learning defi-cits. J Neurosci 2003;23:876–882.

41. Johnson SA, Young C, Olney JW. Isoflurane-induced neuroapop-tosis in the developing brain of nonhypoglycemic mice. J Neu-rosurg Anesthesiol 2008;20:21–28.

42. Gressens P, Dingley J, Plaisant F, et al. Analysis of neuronal,glial, endothelial, axonal and apoptotic markers following mod-erate therapeutic hypothermia and anaesthesia in the devlopingpiglet brain. Brain Pathol 2008;18:10–20.

43. Wilhelm S, Ma D, Maze M, Franks NP. Effects of xenon on invitro and in vivo models of neuronal injury. Anesthesiology 2002;96:1485–1491.

44. Dingley J, Hobbs C, Ferguson J, et al. Xenon/hypothermia neu-roprotection regimes in spontaneously breathing neonatal ratsafter hypoxic-ischemic insult: respiratory and sedative effects.Anesth Analg 2008;106:916–923.

45. Dingley J, Tooley J, Porter H, Thoresen M. Xenon providesshort-term neuroprotection in neonatal rats when administeredafter hypoxia-ischemia. Stroke 2006;37:501–506.

46. Klinger G, Beyene J, Shah P, Perlman M. Do hyperoxaemia andhypocapnia add to the risk of brain injury after intrapartum as-phyxia? Arch Dis Child 2005;90:F49–F52.

47. Thoresen M, Hobbs CE, Wood T, et al. Cooling combined withimmediate or delayed xenon inhalation provides equivalent long-term neuroprotection after neonatal hypoxia-ischemia. J CerebBlood Flow Metab 2009;29:707–714.

48. Dietrich WD, Lin B, Globus MY, et al. Effect of delayed MK-801(dizocilpine) treatment with or without immediate postischemichypothermia on chronic neuronal survival after global forebrainischemia in rats. J Cereb Blood Flow Metab 1995;15:960–968.

49. McAuliffe JJ, Loepke AW, Miles L, et al. Desflurane, isoflurane,and sevoflurane provide limited neuroprotection against neonatalhypoxia-ischemia in a delayed preconditioning paradigm. Anes-thesiology 2009;111:533–546.

Chakkarapani et al: Xenon Enhances Neuroprotection

September, 2010 341