Analysis of Neuronal, Glial, Endothelial, Axonal and Apoptotic Markers Following Moderate Therapeutic Hypothermia and Anesthesia in the Developing Piglet Brain

Abstract

Hypothermia (HT) by whole body (WBC) or selective head cooling (SHC) reduces hypoxic‐ischemic (HI) brain injury; however, whether prolonged hypothermia and/or anesthesia disrupts immature brain development, eg, increases apoptosis, is unknown. Anesthesia increases apoptosis in immature animals. We investigated whether neuroprotective hypothermia and anesthesia disrupts normal brain development. Thirty‐eight pigs <24 h old were randomized between five groups and were killed after 72 h: eighteen received a global hypoxic‐ischemic insult under anesthesia, eight subsequently cooled by SHC with WBC to T_rectal 34.5°C for 24 h, followed by 48 h normothermia (NT) at T_rectal 39.0°C, while 10 remained normothermic. Sixteen underwent anesthetized sham hypoxic‐ischemic, six then following normothermia and 10 following hypothermia protocols. There were four normothermic controls. The hypothermia groups demonstrated significant brain hypothermia. In the hypoxic‐ischemic groups this conferred ~60% neuroprotection reducing histological injury scores in all brain areas. Immunohistochemical/histochemical analyses of neuronal, glial, endothelial, axonal, transcriptional apoptotic markers in areas devoid of histological lesions revealed no hypothermia/normothermia group and differences whether exposed to hypoxic‐ischemic or not. Neither 36‐h anesthesia nor 24‐h hypothermia produced adverse effects at 4‐day survival on a panel of brain maturation/neural death markers in newborn pigs. Longer survival studies are necessary to verify the safety of hypothermia in the developing brain.

INTRODUCTION

Selective head cooling (SHC) combined with mild whole body cooling (WBC) (8) or WBC alone (22) have recently been shown to be effective as neuroprotective approaches for the prevention of neonatal hypoxic‐ischemic (HI) encephalopathy in human full‐term babies. Indeed, experimental studies in both adult and newborn animal models have found that hypothermia (HT), when applied with minimum delay after the insult and for an appropriate degree and duration, significantly reduces hypoxic‐ischemic neuronal cell death (1, 4, 8, 9, 12, 24, 27).

Both pilot studies and randomized trials of cooling infants for 72 h have been completed to test the feasibility and safety of cooling regimes. These studies have not revealed any increased mortality, nor risk of clinical complications including infection, hypotension or hemorrhage (2, 6, 10, 22, 23). However, potential adverse effects of prolonged hypothermia on normal brain development per se are difficult to assess by clinical examination on an 18‐month‐old child. As such, this has not yet been done. In addition, neither the short‐ nor long‐term effects of hypothermia per se on brain development have been examined experimentally.

One of the immediate effects of hypothermia is reduced electrical activity. In newborn rats, transient blockade of electrical activity by N‐methyl‐D‐aspartate receptor antagonists or by GABA‐A receptor agonists has been associated with a dramatic increase in developmental apoptosis (14). In view of this finding, evaluation of the effects of hypothermia on neonatal brain development, both in injured and uninjured animal models, is essential before the extensive use of this neuroprotective technique in human neonates.

The present study, performed in newborn piglets, aimed to determine the effects of 24 h of moderate therapeutic hypothermia per se, in sham experimental animals or after a hypoxic‐ischemic insult, on a panel of markers of brain maturation and neural cell death, including apoptosis.

MATERIAL AND METHODS

Animal preparation.

The protocol was carried out under Home Office license in accordance with UK guidelines. Thirty‐eight, crossbred Landrace/Large White, term‐born piglets of either sex, median age 22 h old (range 9–28) and median weight of 1680 g (range 1230–1950) were collected from the sow within 3 h of commencement of the study.

Eighteen piglets underwent a global hypoxic‐ischemic insult, after which eight were cooled for the first 24 h. The remaining 10 hypoxic‐ischemic animals were kept normothermic at 39°C throughout the procedures. Normal core body temperature for newborn pigs is 39°C (19).

Sixteen pigs were anesthetized, instrumented and had sham exposure to a hypoxic‐ischemic insult (no hypoxia). Of these, six followed the hypothermia protocol and 10 the normothermia (NT) temperature protocol described above with ventilation during full anesthesia for the first 24 h. Four further animals were kept at normothermia and served as juvenile controls. Control animals were hand reared with pig formula milk for 4 days.

Anesthesia, ventilation, vascular access, fluid administration and monitoring were carried out as previously described (27). Briefly, during the study, anesthesia was maintained with halothane 0.7%, N₂O 65%, and oxygen ~35% with ventilation settings being adjusted to keep the end‐tidal carbon dioxide (ETCO₂) between 4.5 and 5.5 kPa, except during the insult when the ETCO₂ was not controlled. After insertion of the monitoring probes, the pig was placed in a prone position and rested for a 60‐minute stabilizing period. The hypoxic‐ischemic pigs were then subjected to a 45‐minute hypoxic‐ischemic insult by lowering the FiO₂ to ~5% O₂ to maintain the EEG voltage ≤7 µV. At the end of the 45‐minute exposure, the low FiO₂ was terminated and the piglets were resuscitated with 30% FiO₂ for a few minutes to restore transcutaneous oxygen saturation to 95%–98% (Masimo Set, San Diego, CA, USA), followed by ventilation with air to avoid hyperoxia during the reoxygenation period.

After the insult, the animals were immediately randomized to either normothermia (T_rectal 39.0°C) or hypothermia (T_rectal 34.5°C + SHC) for 24 h followed by 48 h of normothermia in both groups. Inhalation anesthesia was replaced with intravenous anesthesia (maintenance infusions of propofol; 6 ± 1 mg/kg/h, and fentanyl; 10 µg/kg/h for 24 h) until rewarming began in the hypothermia group. At this point, the anesthesia was stopped and analgesia was provided by intramuscular buprenorphine (20 µg/kg/12 h). Once anesthesia had worn off, those pigs that were able to breath spontaneously were extubated (8/8 hypothermia and 8/10 normothermia).

Mean arterial blood pressure (MABP) was measured continuously from the umbilical artery catheter. During the insult no lower limit for MABP was set, unless it was associated with severe bradycardia (HR < 100) in which case FiO₂ was slightly increased to improve cardiac function. After hypoxic‐ischemic, a MABP <40 mmHg was treated with up to two boluses of 10 mL/kg of 0.9% normal saline, followed by inotropic support as needed (dopamine; range 5–20 µg/kg/minutes plus, if necessary, noradrenaline; range 0.1–1 µg/kg/minutes). Six normothermia and three hypothermia animals required inotropic support, while none of the sham experimental animals did.

Temperature monitoring.

Deep rectal and skin surface (thigh and scalp) temperatures were recorded (400 series probes from Yellow Springs Instrument OH, USA). Intracerebral temperature was measured in the superficial cortex and also from the basal ganglia at a depth of 2.2 cm from the brain surface with microthermocouples (Physitemp Instruments Ltd, Clifton, NJ, USA) calibrated against a certified mercury‐in‐glass thermometer (BS593, Zeal, London, UK) within ±0.1°C over a temperature range 20°C to 40°C. For backup purposes, all temperatures and physiological variables were both computer logged and recorded manually at 30‐minute intervals throughout the 72‐h experiment.

EEG monitoring.

Single channel biparietal (interelectrode distance of 3 cm at P3‐P4 position) amplitude‐integrated EEG was recorded continuously at baseline and during cooling‐rewarming and analyzed using a cerebral function monitor (CFM 6000, Olympic Medical, Seattle, WA, USA).

Temperature control.

Pigs in the experimental and sham normothermia groups were maintained throughout the study period at the desired T_rectal 39 ± 0.2°C under a radiant heater. Pigs in the experimental and sham hypothermia groups were cooled for 24 h using a cooling cap with a network of circulating water channels (Olympic Cool Care System). The cap was fitted around the head and snout. Contact with the neck was avoided. Cooling was started by circulating water through the cap, initially at a temperature of 9°C, with the overhead heater turned off. When the rectal temperature had fallen from 39°C to 34.5°C, overhead heating of the exposed body was restarted, with the head shielded by a heat‐reflecting barrier. The target steady state with a stable T_rectal of 34–35°C was achieved by 50 minutes (range 35–60). The median skin temperature on the scalp under the cap was ~23°C, the brain temperature of the superficial cortex was ~27°C and the deep brain (basal ganglia) temperature ~30°C. After 24 h, rewarming was carried out over 6 h using an overhead heater to ensure steady increase in T_rectal of 0.8°C/h up to 39°C.

Histopathology.

At 72 h post insult, the animals were reintubated, deeply anesthetized with halothane and the brains fixed with 4% phosphate‐buffered formaldehyde by perfusion through the common carotid arteries. Six areas of the brain were examined (neocortex, white matter, hippocampus, cerebellum, basal ganglia and thalamus). The severity of damage in each region was assessed on hematoxylin‐eosin (HE) and icresyl‐violet stained sections and graded with 0.5‐intervals from 0.0–4.0 giving a 9‐step scale (25). A score of zero represented 0% damage, 1.0 represented ≤10% of the area affected, 2.0 (20–30% damaged), 3.0 (40–60% damaged) and 4.0 (>75% of the area damaged) (25, 27). Only cases with a pathology score lower than 3.5 in all anatomical areas were included in subsequent comparative analyses in order to find some uninjured areas for examinations. Accordingly, sections from one animal of the hypoxic‐ischemic‐hypothermia group and three animals from the hypoxic‐ischemic‐normothermia group were not used because the injury was too severe (3.5 or 4.0 in at least one area) with tissue lysis that made it unsuitable for further analysis.

Histochemistry, immunohistochemistry, terminal deoxynucleotidyl transferase‐mediated dUTP nick‐end labeling (TUNEL), and Fluoro–Jade B staining.

Serial histological sections at the level of the parietal cortex including the caudate, thalamus, putamen and globus pallidus nuclei of the basal ganglia and the hippocampus were used for Bodian staining (a marker of axons), immunohistochemistry, TUNEL staining (a marker of fragmented DNA) (Roche, Meylan, France) and Fluoro–Jade B staining (a marker of cell death) (Histo‐Chem Inc., Jefferson, AR, USA). Antibodies used for immunostaining were directed against single strand DNA (Apostain, a marker of fragmented DNA) (1/100, AbCys, Paris, France), neuronal nuclei (NeuN, a marker of neurons) (1/500, mouse monoclonal; Chemicon, Temecula, CA, USA), amyloid precursor protein (APP, a marker of axonal transport) (1/500, rabbit polyclonal, Sigma, St Louis, MI, USA), calbindin protein (CABP, a marker of GABAergic interneurons) (1/2000, rabbit polyclonal, Surant, Bellinzona, Switzerland), vasoactive intestinal peptide (VIP, a marker of some interneurons) (1/500, rabbit polyclonal, DiaSorin, Stillwater, MI, USA), methyl CpG binding protein 2 (MecP2, a transcription repressor) (1/250, rabbit polyclonal, Affinity Bioreagents Inc., Golden, CO, USA), glial fibrillary acidic protein (GFAP, a marker of astrocytes) (1/500, rabbit polyclonal, Dako, Glostrup, Denmark), glucose transporter 1 (GLUT1, a marker of endothelial cells) (1/100, rabbit polyclonal, Biogenesis, Poole, UK), isolectin B4 (a marker of microglia‐macrophages) (1/200, biotinylated lectin, Vector, Burlingame, CA, USA), RCA‐1 lectin (a marker of microglia‐macrophages) (1/2000, biotinylated lectin, Vector), and cleaved caspase‐7 (a marker of apoptosis) (1/200, rabbit polyclonal, Cell Signalling, Beverly, MA, USA).

For immunohistochemistry, deparaffinized sections were microwaved prior to overnight incubation with the primary antibodies. These antibodies were detected using an avidin‐biotin‐horseradish peroxidase kit (Vector), as instructed by the manufacturer. Diaminobenzidine was used as a chromogen.

TUNEL staining was performed using an in situ cell death detection kit as instructed by the manufacturer. In brief, sections were deparaffinized, treated for 20 minutes at 37°C with proteinase K, and incubated for 2 minutes on ice with 0.1% Triton X‐100. DNA strand breaks were identified by using terminal deoxynucleotidyl transferase for 60 minutes at 37°C, to label free 3′‐OH termini with fluorescein‐labeled nucleotides. Incorporated nucleotides were detected using an anti‐fluorescein antibody conjugated to alkaline phosphatase, with nitroblue tetrazolium and 5‐bromo‐4‐chloro‐3‐indolyl phosphate toluidonium salt as the substrates.

Fluoro–Jade staining was performed on deparaffinized sections as instructed by the manufacturer. Following ethanol and 0.06% potassium permanganate pre‐treatments, sections were incubated for 30 minutes with 0.001% Fluoro–Jade B solution. Sections were then rinsed, mounted and examined under an epifluorescence microscope using a FITC filter.

Analysis of data.

For each marker, at least three sections of each brain were stained in two successive batches. To avoid regional and experimental variations in labeling intensity, for each marker, sections from the different experimental groups were treated simultaneously. In cases with a pathology score lower than 3.5 in all anatomical areas, quantitative and qualitative analyses were performed in areas which did not show any pathological changes in adjacent sections stained with cresyl violet and HE. Experiments where maximum injury was seen in all areas were excluded from analysis (three normothermia and one hypothermia were excluded, leaving five normothermia and nine hypothermia in the hypoxic‐ischemic experimental group). All markers were analyzed qualitatively at the levels of the parietal cortex, the underlying white matter, the basal ganglia and the hippocampus. Furthermore, cells, nuclei or structures labeled with NeuN, APP, CABP, VIP, MecP2, GFAP, isolectin B4, RCA, GLUT1, cleaved caspase‐7, TUNEL, Fluoro–Jade and Apostain were quantified at the level of the parietal cortex. Six to 18 fields were analyzed in each experimental group for each animal and for each marker. The surface area of each field for each marker is given in the figures. Qualitative and quantitative analyses were performed by an observer unaware of the experimental groups.

Statistical analysis.

Quantitative data were expressed as the mean ± SEM values for each treatment group. Results were compared using ANOVA with Bonferroni’s multiple comparison of means test. Differences were considered statistically significant when P < 0.05.

RESULTS

We have previously shown that a short duration (3 h) of hypothermia starting immediately after hypoxic‐ischemic is neuroprotective in the case of mild, but not severe insults (12). However, SHC combined with WBC for 24 h, followed by 48 h normothermic survival, offers significant protection after hypoxic‐ischemic; not only after a mild insult, but also in the case of more severe injuries and seizures (27). Figure 1 shows the regional brain injury in the normothermia and hypothermia groups. Hypothermia offered significant neuroprotection, as previously shown (25). In this model, we have not seen any differences between genders in the response to a hypoxic‐ischemic insult, nor in the protective effect of hypothermia (27). Figure 2 shows typical examples of lesions of different severity from neocortex, white matter, basal ganglia, hippocampus and cerebellum.

Figure 1.

The hypothermia group (n = 8), cooled for 24 h by a combination of mild whole body hypothermia and selective head cooling, was compared with normothermia (n = 10) (27). A validated 9‐step scoring system was used ranging from 0.0 (no injury) to 4.0 (75%–100%) injury (25). Values are median with standard error around the median.

Figure 2.

Photomicrographs showing regional histopathological (hematoxylin‐eosin stain) features typical of grade 0.0 (left column), grade 2.0 (middle column) and grade 3.5 to 4.0 (right column) (25). Survival was 72 h after the hypoxic‐ischemic insult. Images from neocortex (A–C), periventricular (D) and subcortical (E,F) white matter, CA1 area in the hippocampus (G–I), basal ganglia (J–L) and cerebellum (M–O) are shown. Infarcts, identified as pale areas with extensive cell necrosis and vacuolation of the neuropil, are larger in higher scoring grades. Arrows in (H) point to apoptotic neurons with homogenous eosinophilic cytoplasm and pyknotic nuclei. Arrows in (M), (N) and (O) point to Purkinje cells. The magnification of each image is shown in the lower right.

Using areas of parietal cortex which did not exhibit any pathological sign (and were thus scored as 0) in cresyl violet or HE staining performed on adjacent sections, quantification of cortical cells labeled with anti‐NeuN (3, 4), anti‐MecP2 (3, 4), anti‐CABP (3, 5), anti‐VIP (3, 5), anti‐GFAP (3, 6), anti‐GLUT1 (3, 6), anti‐cleaved caspase‐7 (7, 8) or TUNEL (Figure 7) revealed no significant differences between the post‐hypoxic‐ischemic insult normothermia and hypothermia groups, nor the sham hypoxic‐ischemic hypothermia or normothermia groups. Also these results were similar to those found in juvenile controls that had not been subjected to either a hypoxic‐ischemic insult or the anesthetic protocol (at least 6 h of inhalation followed by 30 h of intravenous anesthesia). Staining adjacent sections with isolectin B4, RCA or Fluoro–Jade did not label cells in any tested groups (data not shown). Furthermore, qualitative assessment of all of these markers in the basal ganglia and hippocampus revealed no detectable difference between the experimental groups (data not shown).

Figure 3.

Quantification of neocortical cells labeled with anti‐NeuN (A), anti‐MecP2 (B), anti‐CABP (C), anti‐VIP (D), anti‐GFAP (E), or anti‐GLUT1 (F) in the different experimental groups. NT = normothermia; HT = hypothermia.

Figure 4.

Typical examples of neocortical labeling with anti‐NeuN (A,B) and anti‐Mecp2 (C,D) in NT_sham (A,C) and HT (B,D) piglets. Bar = 20 µm. NT = normothermia; HT = hypothermia.

Figure 5.

Typical examples of neocortical labeling with anti‐CABP (A,B) and anti‐VIP (C,D) in NT_sham (A,C) and HT (B,D) piglets. Arrows point to examples of labeled cells (C,D). Bar = 20 µm. NT = normothermia; HT = hypothermia.

Figure 6.

Typical examples of neocortical labeling with anti‐GFAP (A,B) and anti‐GLUT1 (C,D) in NT_sham (A,C) and HT (B,D) piglets. Bar = 20 µm. NT = normothermia; HT = hypothermia.

Figure 7.

Quantification of neocortical cells labeled with anti‐cleaved caspase‐7 (A), TUNEL staining (B) or Apostain (C) in the different experimental groups. NT = normothermia; HT = hypothermia.

Figure 8.

Typical examples of neocortical labeling with anti‐cleaved caspase‐7 in NT_sham (A) and HT (B) piglets. Arrows point to examples of labeled cells. Bar = 20 µm. NT = normothermia; HT = hypothermia.

Qualitative analysis of fiber tracts in the subcortical white matter stained with APP or Bodian exhibited no differences between the different groups (Figure 9) in brain areas without any pathological changes in the subcortical white matter in adjacent sections stained with cresyl violet or HE.

Figure 9.

Typical examples of periventricular white matter labeling with anti‐APP in NT_sham (A) and HT (B) piglets. Bar = 20 µm. NT = normothermia; HT = hypothermia.

DISCUSSION

The present study, which used a panel of markers of neural cell differentiation and death, revealed no detrimental effects on the brains of 24 h of hypothermia in newborn piglets that had survived for at least 72 h. These piglets were either juvenile controls, sham controls subjected to hypothermia, or animals that were kept hypothermic as a neuroprotective strategy after a global hypoxic‐ischemic insult. The piglets in these hypothermia groups were compared with the corresponding groups kept at normal piglet body temperature (39°C).

Relevance of the piglet model to human term neonates.

Our piglet model is the only global hypoxic‐ischemic model which exhibits a pattern of neural injury similar to that in the newborn human infant after a hypoxic‐ischemic insult, as well as development of single or multiorgan failure (12, 21, 24, 25, 27). Like the encephalopathic newborn, this global model develops post‐hypoxic seizures and allows post‐insult survival with awake neurological testing that correlates with the pathological outcome (25).

Short (3 h) (12) and long (24 h) (27) post‐hypoxic hypothermia have been shown to be neuroprotective following mild or more severe injury respectively in this model.

The maturation of the piglet brain is similar to that of the term infant with respect to the timing of the “growth spurt” (5), as are the cardiovascular responses to CO₂ and blood pressure changes (11).

The effect of anesthesia.

In the immature rat, anesthesia was found to induce apoptosis (15). In this study, two of the groups underwent a sham insult under full anesthesia, followed by either 24 h of hypothermia + 48 h normothermia or 72 h of normothermia. There was no increase in the number of apoptotic cells as assessed by our three different methods (TUNEL, Apostain and Fluoro‐Jade), in either of these groups as compared with juvenile controls. Hence, we did not reproduce the deleterious effect of anesthesia in the newborn pig.

None of the markers used in this study revealed any cellular injury in regions where no injury was identified on HE staining in any of the controls or hypoxic‐ischemic animals, whether or not they were cooled and/or anesthetized. Other workers (12, 18) have observed no detectable neurotoxic effects in sheep fetuses after maternal general anesthesia. These apparent discrepancies may be species or anesthesia dependent, as different anesthesia protocols were used in the three studies.

It is worth considering whether it is essential for hypothermia to be combined with anesthesia, or at least sedation, to be protective. Most large animal cooling studies and all adult human studies (3, 13) have taken place with concurrent anesthesia or sedation. We have subjected our piglet model to 24 h of post‐insult hypothermia without sedation or anesthesia (26). Figure 10 shows that hypothermia did not exert any protection at all under these conditions, quite contrary to the results shown in Figure 1, where the animals were sedated/anesthetized under the same protocol. The cortisol levels during unanesthetized hypothermia were three times that during hypothermic anesthesia and we hypothesize that the stress of being awake and cold has a deleterious effect on brain recovery. One study by Gunn et al in fetal sheep demonstrated hypothermic neuroprotection without sedation (9). However, the pregnant state exerts less stress on the fetus as compared with the situation after birth. The clinical data from infants does not fully answer the question. The infants in the neonatal cooling studies (showing benefit) were sedated as clinically indicated (8, 22) and data on sedation has not been presented.

Figure 10.

Data from previous work in the piglet model where 24 h of hypothermia (WBC) without concomitant sedation or anesthesia after a hypoxic‐ischemic insult was used (26). There is no protection from 24 h of unsedated hypothermia after a hypoxic‐ischemic insult. Values are mean and SEM.

It is also worth considering whether anesthetics might be neuroprotective in their own right. There is evidence that some anesthetics, such as isoflurane, can act as a preconditioning agent on the brain (28). Post‐insult inhaled anesthetics such as isoflurane, as well as some intravenous agents such as barbiturates, may be neuroprotective if the ischemic insult is mild (16, 17). However, with moderate/severe insults, sustained protection is not generally evident and the neuroprotective potency of anesthetics such as propofol, if any, still remains undefined (7).

Impact of hypothermia on the developing brain.

Little is known about how to equate the duration of an intervention during a period of rapid development between human and animal neonates with markedly different speeds of maturation. As the pig matures at least 20 times faster than a human. As such, is 24 h of hypothermia after birth relatively longer than the 72 h used in the human trials (8, 22)? Similarly, what is the human equivalent of 4 days survival in the neonatal pig? Despite pig brain maturation at birth being, in most respects, similar to that of the human at term, the kidney and the liver are less mature in the former than in the latter.

As there are important differences between species, it is essential to test the effect of hypothermia in several species of different maturation before assuming that effects observed in any particular one will also be found in humans.

The markers used in this study allow examination of important cell populations and features of the developing brain such as total neuronal population (NeuN), GABAergic (CABP) and VIPergic (VIP) interneurons, reactive microglia (lectin and RCA), astrogliosis (GFAP), microvasculature (GLUT1), axons (APP and Bodian), neural cell death (TUNEL, Apostain, cleaved caspase‐7 and Fluoro–Jade), Some markers, such as RCA, lectin and Fluoro–Jade, were negative in areas without detectable lesions in adjacent cresyl‐violet‐stained sections. However, the same markers strongly labeled areas with histological lesions (data not shown), used as positive controls.

The present data provide encouraging evidence that the use of mild hypothermia for neuroprotection of the brain of term newborns at risk of developing hypoxic‐ischemic brain lesions is a safe procedure. However, the present data require further extension and validation because of the relatively short duration of the survival period. Further experimental studies seem mandatory to establish beyond doubt the safety of hypothermia as a neuroprotective strategy in human term newborns. These may be summarized as: (i) to test the effects of hypothermia on a larger panel of markers for different proteins and/or mRNAs critical for brain development and maturation, including receptors or transcription factors (in this regard, the pig model has some limitations because relatively few porcine antibodies are available); (ii) to test these markers at later time points in order to identify possible long‐term effects of hypothermia on brain development; (iii) to test the effects of hypothermia on other indicators of brain maturation such as dendritic spine plasticity (which has been shown to be inhibited in vitro by deep hypothermia) (20), electrophysiological parameters, learning and cognitive behaviors; and (iv) to confirm the data obtained in piglets in other animals.

In summary, the present study provides new experimental evidence that is consistent with the safety of using neuroprotective hypothermia. In addition, it shows that full anesthesia, for a minimum of 36 h, caused no injury examined after 4 days in a term equivalent newborn hypoxic‐ischemic model. Nevertheless, more experimental studies, in particular with long‐term survival, are desirable to establish the full safety profile of this procedure when applied to a developing brain.