Aeromedical evacuation-relevant hypobaria following traumatic brain injury in rats contributes to cerebral blood flow depression,altered neurochemistry and increased neuroinflammation

Rat model of controlled cortical impact TBI

All animal procedures were conducted in accordance with the University of Maryland School of Medicine (IACUC #0317005) and the US Army and the U.S. Air Force Surgeon General’s Office of Research Oversight and Compliance (FWR-2017-0018A) according to the Guide for the Care and Use of Laboratory Animals (National Research Council, 2011). Reporting of animal experiments herein are in compliance with Animal Research: Reporting in Vivo Experiments (ARRIVE) guidelines. Rats used in these experiments were double housed prior to injury and single housed following experimental procedures to prevent additional trauma to incisions commonly inflicted by cage mates. One-inch wooden blocks for chewing and manipulation were provided as environmental enrichment during solitary housing. Adult male Sprague-Dawley rats (250–350 grams, n = 35) were subjected to a single left parietal controlled cortical impact (CCI) injury or sham injury (craniotomy). ¹⁴ TBI was performed using the CCI device (ImpactOne, Leica). Rats were anesthetized with 4% isoflurane, positioned in a stereotactic frame, then maintained under 1.5–2.5% isoflurane for the duration of surgical preparation. The skull was exposed via a midline incision and a high-speed dental drill (Henry Schein, Melville, NY) was used to perform a left-sided craniotomy above parietal cortex centered 3.5 mm posterior and 4 mm lateral to bregma. A 5 mm beveled impactor tip was accelerated to 5 m/sec with a vertical deformation depth of 1.5 mm and impact duration of 50 ms, consistent with mild injury. ¹⁵ The excised bone was replaced and sealed with dental acrylic. After closing the scalp incision with Ethilon sutures, anesthesia was discontinued. Buprenorphine (0.01 mg/kg) was administered for analgesia immediately prior to surgery and then for three days post-operatively during which time animals were monitored for weight loss, dehydration and lethargy.

Magnetic resonance compatible hypobaric chamber

The hypobaric chamber was constructed around the standard rat MRI bed as it contains the nose cone and other components essential for conducting imaging experiments. A polycarbonate cylinder (4 ft in length, 6.35 mm thick, 82 mm diameter) was used as the chamber body, sealed with removable end caps custom machined with multiple ports (Suppl Figure 1). On the end cap closest to the animals’ head, provisions were made to allow for a 4-channel phased array receiver coil and for providing gas anesthesia. On the other end, ports allowed passage of wire leads (temperature probe and respiration monitor) and tubing (warm water circulation, tubing for vacuum generation, and anesthesia gas circulation). Once the instrumented rat was placed into the animal bed and placed inside the hypobaric chamber, the chamber was sealed by capping both ends. An 86 mm linear volume array coil was then slipped over the plastic chamber and positioned at the center of the rat head with the 4-channel surface coil already in place. Simulated AERH initiated at 24 hpi was maintained for a total of six hours, including a separate 20 minute pressurization and depressurization adjustment period (Suppl Figure 2). Chamber pressure was controlled by adjusting air flow from a constant vacuum generator until 10.9 psi was reached simulating air transport conditions (hypobaria, equivalent to 8000 ft altitude). No vacuum was applied to the chamber for experiments conducted under ground transport conditions.

OPEN IN VIEWER

Figure 1.

Representative brain MRI images across four bregma levels and four timepoints. (a) Coronal T2-weighted MR images from a rat following craniotomy alone (b) coronal T2- weighted MR images showing evolution of injury following controlled cortical impact BL: baseline, BT: before transport, DT: during transport, AT: after transport.

OPEN IN VIEWER

Figure 2.

Representative in vivo localized 1H MR spectra taken from the ipsilateral cortex (white box) (a) before injury and (b) during transport. (c) Coronal T2-weighted MR images (left column) and corresponding CBF maps (right column) showing six regions of interest for CBF maps. GABA: γ-aminobutyric acid, Glu: glutamate, Gln: glutamine, Glx: Glu + Gln, GSH: glutathione, Ins: myo-inositol, Lac: lactate, NAA: N-acetyl aspartate, Tau: taurine, tCr: creatine + phosphocreatine, tCho: glycerophosphocholine + phosphocholine, MM: macromolecules. BL: baseline, BT: before transport, DT: during transport, AT: after transport. 1: ipsilateral cortex. 2: contralateral cortex. 3: ipsilateral hippocampus. 4: contralateral hippocampus. 5: ipsilateral thalamus. 6: contralateral thalamus.

MRI and MRS data acquisition

All MRI and MRS studies were performed at the Center for Translational Research in Imaging at the University of Maryland School of Medicine (Baltimore) using a Bruker Biospec 7.0 Tesla 30 cm horizontal bore scanner (Bruker Biospin MRI, Ettlingen, Germany), which was interfaced to a Bruker Paravision 6.0 console for the acquisition of MRI and MRS data. A Bruker 86 mm circular-polarized RF volume coil worked as a transmitter, and Bruker 4-channel array surface coil as a receiver.

Coronal T₂-weighted Bruker rapid imaging with refocused echoes (RARE) images were obtained (spin-echo, Repetition time (TR)/echo time (TE) 3000/22 ms, two averages, RARE factor 4, field of view (FOV) 40 × 40 mm², image matrix 256 ×  256, number of slices 16, slice thickness 1 mm, with an acquisition time 6 min 24 s). These images served to verify positioning over the center of the CCI and to determine abnormalities in T₂-weighted signal intensity within the slice chosen for subsequent measurements of perfusion and spin-lattice relaxation time (Figure 1(a) and (b)).

CBF was measured using continuous arterial spin labeled (CASL) MRI technique with a spin-echo read-out, a TR/TE of 5000/26.72 ms, FOV 40 × 40 mm², image matrix 64 × 64, slice thickness 2 mm, in the coronal plane where 5 labeled images and 5 control images with an acquisition time 50 sec on a slice that covered the impacted cortex, hippocampus, and thalamus (Figure 2(c)).^16,17 CBF was calculated from:

CBF = λ × ( T 1obs × 2α ) − 1 × ( [ MC − ML ] / MC )

(1)

M ( TR ) = M 0 [ 1 − A exp ( − TR / T 1obs ) ]

(2)

Proton spectra were obtained from both the left and right cortex (3.5 × 1.5 × 4 mm³) using a proton short-TE Point-RESolved Spectroscopy (PRESS) pulse sequence, TR/TE = 2500/10 ms, and 400 averages, respectively. ¹⁹ Sample MR spectrums collected from the cortical region showing signals associated with multiple neurochemicals are shown in Figure 2(a) and (b). The signal intensity, or area under the peak, was calculated and corresponded to a concentration in parts per million (ppm) for each specific neurochemical. The unsuppressed water signal from each of the prescribed voxels was obtained to serve as a reference for determining the specific neurochemical concentrations.

MRI/MRS data collection from rats

Animals were anesthetized (2–5% isoflurane) in an induction chamber, then placed prone on an animal bed with designated O₂ level mixed with medical air administration at 0.8 L/min. The head was fixed with a bite bar and a pair of ear pins. The 4-channel radio frequency receiver coil was then centered and fixed over the head. Imaging was initiated while the anesthesia was maintained at 1–3% isoflurane via nosecone throughout the rest of the imaging session. An MR-compatible system was used to monitor respiration rate, and body temperature was maintained at 35–37.5°C using a circulating warm water heater. Pulse oximetry measurements made under anesthesia during scanning procedures indicated that SpO₂ was maintained at 95–99% saturation for all animal groups.

Magnetic resonance scans were collected at the following time points: Baseline 48 hr before surgery (BL), 24 hr after surgical injury immediately before transport (BT), during transport (DT), and 14 days after transport (AT).

Baseline scans and those collected before and after transport were performed under simulated “ground transport” (normobaria at sea level) and normoxic (30% O₂) conditions. Scans conducted during transport were made under a combination of simulated “ground transport” (normobaria at sea level) or air transport (hypobaria equivalent to 8000 ft) barometric pressure and normoxic (30% O₂ with normobaria and 40% O₂ with hypobaria) or hyperoxic (100% O₂) conditions. This level of supplemental O₂ was delivered to prevent hypoxia and was necessary to offset the displacement of O₂ observed with long durations of isoflurane anesthesia and to account for the reduced O₂ availability at increasing altitudes.

MRI/MRS data processing and analysis

Pixel-by-pixel maps of CBF generated from parent images of the perfusion studies using equation (1) and maps of T_1obs generated from the series of variable TR spin-echo were programed using equation (2) in Matlab. Manually drawn regions of interest (ROIs) were initially defined within left (ipsilateral) and right (contralateral) cortex, hippocampus, and thalamus on the control images from perfusion experiments (Figure 2(c)) by assignments from a rat brain atlas. ²⁰ Due to TBI induced brain damage, and resulting edema, images showed structural differences in injured tissue. These differences have the potential to vary from time point to time point in the same animal, especially around the three ipsilateral ROIs. The degree of brain tissue damage might also be different among different animals. Given these confounders, we manually drew the ROIs in each animal at each time point, respectively. After the neuroanatomical region for quantification was identified, a smaller ROI was manually drawn inside the boundaries of that specific nuclei. For this reason, even if the ROI location was slightly mismatched from one time point to another, it was still within the target anatomical region. To verify whether reliability of results, we varied the sizes and locations of the ROIs in the same animal with consistent results.

CBF was calculated for each ROI using equation (1). A spatially constant value of 0.9 ml/g was assumed for the blood-brain partition coefficient λ.²¹ The labeling efficiency, α was set to 0.7 for all CBF calculations. ²²

Quantification of the MRS was based on frequency domain analysis using a “Linear Combination of Model spectra” (LCModel). ²³ Absolute concentrations were estimated with the LCModel automatic procedure (version 6.3-0 G). Only those neurochemicals that passed the Cramer-Rao lower bound of ≤30% were considered for further analysis. The following neurochemicals were identified gamma amino butyric acid (GABA), glutamine (Gln), glutamate (Glu), glutathione (GSH), glutamate plus glutamine (Glx), lactate (Lac), N-acetyl aspartate (NAA), Inositol (Ins), taurine (Tau), total creatine (tCr), total choline (tCho), and macromolecules (MM).

Cortical areas on T₂ images appearing bright white were classified as “hyperintensities” (HI) and areas appearing matte gray were classified as “healthy” (Figure 1(a) and (b)). Areas quantified from 1 mm thick sections were summed to calculate cortical volumes (mm³) at each of the four timepoints using ImageJ.

Immunohistochemical evaluation of neuroinflammation

Brains were perfusion fixed with 4% paraformaldehyde at 14 dpi and immunostained using antibodies against ED1 (mouse Anti-ED1, 1:100,000, Bio-Rad), a marker of inflammatory activation of macrophages and microglia. ⁶

Seven ED1 immunolabelled brain sections (40 µm thick, 960 µm apart) were used for stereologic quantification of cortical lesion volume (mm³), the first corresponding to the epicenter of cortical injury, then 3 sections both rostral and caudal to that section. Volume/area were measured according to the Cavelieri method. ²⁴ Analysis was performed using a Nikon E800 motorized microscope under 10x magnification by a microscopist blinded to animal treatment group. A 25 × 25 µm grid was generated and placed by StereoInvestigator software (Microbrightfield Biosciences, Vermont, USA). The impact-induced necrotic divot was characterized by missing or fenestrated tissue where dorsal borders were estimated based on cortical thickness present on the contralateral hemisphere of the same section. Tissue surrounding the divot demonstrating an abundance of ED1 positive cells was classified as ‘penumbra’. Necrotic divot and inflammatory penumbral volumes were combined and expressed as the total lesion volume.

Cerebral blood flow (CBF)

CBF maps covered ipsilateral and contralateral cortex, hippocampus, and thalamus (Figure 2(c)). Low signal regions indicate decreased CBF values. A summary of group means and standard deviations for regional blood flow measurements are presented in Supplemental Table 1. Craniotomy and TBI alone resulted in an acute reduction in CBF at 24 hr after surgical injury (BT vs BL) to the cortex and hippocampus underlying the site of cortical impact regions (Figure 3(a) and (b)).

OPEN IN VIEWER

Figure 3.

Box-and-Whisker plots of post injury CBF. (a) ipsilateral cortex. (b) ipsilateral hippocampus and (c) ipsilateral thalamus. Effects of Surgical injury: solid line; Effects of hypobaria: bold solid line. Effects of hyperoxia: dotted line. Asterisks indicate significance compared to groups own Pre-Injury Baseline values. All lines and asterisks indicate p < .05.

Post-injury hypobaria exposure resulted in lower cerebral blood flow to the bilateral thalamus during flight (DT) compared to groups maintained under normobaria (Figures 3(c) and 4(c)). Although a trend towards lower CBF was also found during transport in the ipsilateral hippocampus of all groups exposed to hypobaria, this decrease was only significant between sham groups, suggesting that hypobaria in general causes a reduction in regional CBF regardless of injury (Figure 3(b)).

OPEN IN VIEWER

Figure 4.

Box-and-Whisker plots of post injury CBFs. (a) contralateral cortex. (b) contralateral hippocampus and (c) contralateral thalamus. Effects of Surgical injury: solid line; Effects of hypobaria: bold solid line; Effects of hyperoxia: dotted line. Asterisks indicate significance compared to groups own Pre-Injury Baseline values. All lines and asterisks indicate p < .05.

In the cortex, acute differences in blood flow due to AERH were detected in contralateral, but not ipsilateral, hemispheres at 24 hr after impact (Figure 4(a)). By 14 dpi, brain injured rats exposed to hypobaria under 100% O₂ had lower cortical blood flow (Figure 3(a)) to the injured cortex compared to those maintained at sea level (NB/HYP vs HB/HYP). Rats exposed to HB/HYP also had lower levels of blood flow to the ipsilateral cortex (Figure 3(a)) and bilateral thalamus (Figures 3(c) and 4(c)) at 14 dpi compared to sham groups transported under hypobaria (HB/HYP vs SHAM HB). There were no significant effects of hyperoxia alone on CBF detected at acute or delayed timepoints.

Cortical pathology

General pathophysiological processes including vascular disruption, infarction, hemorrhage, inflammatory cellular infiltration, edema, and gliosis were observed in and around the lesion (Figures 1 and 5), as has been observed by us and others using the CCI model.^15,25,26 Although craniotomy alone resulted in minor localized injury, CCI induced a distinct but heterogeneous area of hyperintensity on T₂-weighted MR images (Figure 5(b) and (c)) at the impact site at 24 hr. Quantification of MR volumes from T₂ images (Figure 5(e)) demonstrated that although the lesion was partially resolved, it persisted at least 14 days after injury compared to sham injured animals. Sham, but not injured, rats exposed to AERH had higher cortical MR hyperintensity volumes during flight. Sham surgery (Figure 5(a)) produced a slight deformation (white arrow) and a localized inflammatory reaction (black arrow), visualized through T₂ MR images and ED1 immunolabeling of macrophages.

OPEN IN VIEWER

Figure 5.

In vivo and ex vivo quantification of cortical pathology. Representative images of ED1 immunolabeled cortex at 14 days after (a) Sham with NB transport (b) CCI with NB/NOR transport or (c) CCI with HB/NOR transport. (d) Inflammatory penumbra quantification from ED1 immunostaining and (e) hyperintensity volume quantification from in vivo MR images. Effects of Surgical injury: solid line; Effects of hypobaria: bold solid line; Effects of hyperoxia: dotted line; Asterisks indicate significance compared to groups own Pre-Injury Baseline values. All lines and asterisks indicate p < .05.

CCI resulted in an ipsilateral necrotic divot along with a sizable ED1+ inflammatory penumbra (Figure 5(d)) at 14 days post injury. The necrotic divot and ED1+ inflammatory penumbra corresponds to the hyperintensity area on T₂-weighted MR images. There was a moderate positive correlation between MR hyperintensity volumes measured during transport and subsequent ED1 histopathology measured at 14 dpi, r(32) = .52, p<.001. Linear regression analysis showed this relationship was significant (p = .002). This moderately positive relationship remained significant up to 14 days after injury, r(32) = .65, p < .001.

Following cortical injury, hypobaria exposure resulted in higher ED1+ inflammatory penumbra volumes (Figure 5(d)) without increasing the cortical hyperintensity volume (Figure 5(b) and (c) white arrow) quantified from T₂ weighted images (Figure 5(e)). We observed no significant effects of hyperoxia on MR volumes or cortical inflammation (ED1) at the delayed timepoint of 14 dpi.

Cortical neurochemicals

¹H magnetic resonance spectra (Figure 2(a) and (b)) demonstrated excellent spectral resolution and sensitivity at each timepoint measured. Cortical neurochemicals were compared from single-voxel spectroscopy taken near the location of the impact and a similar location contralateral to the lesion. A summary of group means and standard deviations for all cortical neurochemicals measured are presented in Supplemental Tables 2 (ipsilateral) and Table 3 (contralateral). Cortical Lac (Supplemental Figure 3) and Gln increased due to surgical injury while concentrations of GABA, GSH, GPC+PCh, NAA, Ins, Tau, tCr, and Glu were reduced at 24 hpi (Figure 6(a) to (d) BL vs BT). Of these metabolites, all but GABA, Glu, NAA, and tCr recovered to baseline levels by 14 dpi (Figure 6). Surgical injury alone did not alter neurochemical levels measured from contralateral cortex at 24 hr or 14 dpi (Figure 7).

OPEN IN VIEWER

Figure 6.

Box-and-Whisker plots of post injury ipsilateral biochemical concentrations. (a) GABA. (b) Glu. (c) tCr and (d) NAA. Effects of Surgical injury: solid line; Effects of hypobaria: bold solid line; Effects of hyperoxia: dotted line. Asterisks indicate significance compared to groups own Pre-Injury Baseline values. All lines and asterisks indicate p < .05.

OPEN IN VIEWER

Figure 7.

Box-and-Whisker plots of post injury contralateral biochemical concentrations. (a) GABA. (b) Glu. (c) tCr and (d) NAA. Effects of Surgical injury: solid line; Effects of hypobaria: bold solid line; Effects of hyperoxia: dotted line. Asterisks indicate significance compared to groups own Pre-Injury Baseline values. All lines and asterisks indicate p < .05.

Following cortical injury, hypobaria exposure resulted in higher levels of total creatine (Figure 6(c)) during flight in the ipsilateral hemisphere compared to injured groups maintained under sea level conditions (NB/NOR vs HB/NOR). Although an acute effect of hypobaria was not detected during flight, injured rats exposed to hypobaria had lower levels of GABA (Figure 7(a)) at 14 dpi compared to injured groups maintained under normobaria (NB/NOR vs HB/NOR).

Following injury, no neurochemical differences due to hyperoxia alone were detected. However, rats exposed to the combination of hyperoxia and hypobaria had lower concentrations of tCr (Figure 6(c)) and Glu (Figure 7(b)) during flight (HB/NOR vs HB/HYP).

TBI reduces CBF, alters neurochemicals, and increases inflammation in the cortex

In the current study, a decline in CBF to the cortex and hippocampus was observed at one day after injury. Under basal conditions, CBF is maintained relatively constant over a wide range of physiological conditions and changes in blood pressure. After TBI, autoregulation of CBF is often compromised, thus reducing CBF.^{27 –29} These TBI-induced CBF disturbances are well studied both in experimental animal models and clinically.^{22,27 –40} Our results are consistent with previous studies showing that regional differences of cortical CBF are heterogeneously affected. ³²

T₂-weighted MR images successfully identify hyperintensities indicative of cortical lesions following CCI. ³² In our study, CCI resulted in an acute increase of MR-based cortical hyperintensity volumes. These edematous volumes partially resolved by 14 days when they were accompanied by increased pathology indicated by cortical inflammation. Further analysis demonstrated that acute T₂ hyperintensity volumes exhibited a moderate correlation with delayed neuropathology obtained with ED1 immunolabeling. The finding that acute in vivo MRI predicts delayed histopathology suggests that with earlier detection, treatment of neuroinflammation and improved long-term recovery may be possible. ⁴¹ Following TBI, we observed an acute reduction in the neurotransmitters glutamate and GABA indicating impaired neurotransmission and a reduction in tCr suggesting oxidative phosphorylation may be impaired. Fluctuations in tCr observed after TBI, hypobaria, and exposure to hypobaria under hyperoxia could contribute to ongoing brain tissue necrosis and inflammation. Lactate was detected within 24 hr of injury as reported by us and others.^{1,11,42,43 1}H MRS results obtained from a similar model of rat CCI that also detected an acute decline in NAA and tCr with partial recovery at one week after injury. However, we failed to observe normalization of NAA by two weeks post-injury, which may be linked to impaired NAA synthesis in damaged mitochondria.^{10,11,44 –46}

Impact of post-injury aeromedical evacuation relevant hypobaria

We found that AERH exposure after TBI resulted in lower CBF, a worsening of cortical inflammation and a partial restoration of injury-induced reductions of cortical creatine. Specifically, one exposure to AERH within 24 hr after injury worsened delayed cortical neuroinflammation, as observed with ED1 immunostaining, beyond that which is seen following TBI alone (Figure 5(d)). These results are consistent with the growing number of studies providing evidence for hypobaria acting as a “second hit” following TBI through worsening of inflammation in the brain and other organs, e.g., the gut, kidney, and lung.^6,47,48

Changes in CBF resulting from hypobaria have been reported in numerous human and animal studies, but the direction of change has not been consistent.^{3,49 –51} The apparently conflicting results may be explained, in part, by the substantial variability in duration and extent of hypobaria, arterial O₂ and carbon dioxide tension during exposure, and the number of previous exposures to hypobaria. Oxygen availability decreases as altitude increases, often resulting in hypobaric hypoxia, rendering it challenging to determine effects of hypobaria independent of inspired O₂ partial pressures. Several studies reported an increase in CBF at high altitude and were conducted under conditions where the reduced partial pressure of O₂ resulted in hypoxia.^49,50 In the current study, hypoxia as a factor was eliminated by simulating normoxia under both normobaria and hypobaria through the delivery of 30 and 40% supplemental O₂ respectively. Our results support studies reporting acute reductions in global CBF after hypobaria exposure compared to those maintained at sea level.^51,52 Using a similar arterial spin labeling technique, Liu et al reported lower CBF in healthy human volunteers after six hr at high altitude compared to their baseline CBFs measured at sea level. ⁵² There is also precedence for reduced CBF in animal models treated with AERH after brain injury where the level of supplemental O₂ is controlled to prevent hypoxia. Scultetus et al demonstrated this reduction in CBF in a swine model of fluid percussion brain injury when injury was followed by four hr of normoxic (O₂ saturation ≤ 92%) AERH identical to the levels used in our study. ³

The partial pressure of O₂ in the blood is a potent regulator of CBF. Hypoxia induces vasodilation of large cerebral vessels including the cerebral and basolateral arteries that supply the thalamus and hippocampus. Likewise, hyperoxia decreases CBF in part by inducing vasoconstriction of these same large vessels. Smaller arterioles surrounded by cerebrospinal fluid within the subarachnoid space have also been implicated in the regulation of CBF via sensitivity to the partial pressures of O₂ and CO₂ in the blood as well as by the local CSF environment which is determined by neurochemical conditions. In the current study, hypobaria exposure partially restored cortical tCr levels and resulted in the lowest cortical GABA and CBF in two of three brain regions measured, independent of injury status or level of O₂ delivered. This resulting hypoperfusion cannot be simply explained by O₂ availability as hyperoxia alone was not a sufficient stimulus to decrease CBF under normobaric pressures. Although the mechanism remains unknown, the possibility exists that regional differences in CBF following AERH may depend on the interaction of several simultaneous factors including cerebrovascular responses to changes in pressure, changes in the local CSF environment and inspired O₂ partial pressures.

Impact of post-injury aeromedical evacuation relevant hypobaria combined with hyperoxia

The combination of hypobaria and hyperoxia reversed the acute restoration of total creatine, and resulted in the lowest cortical and thalamic blood flow measured after simulated air transport.

All cortical metabolites measured had an acute response (24 hpi) to surgical craniotomy with or without cortical impact. The majority of neurochemicals were not further impacted by subsequent exposure to hypobaria or hyperoxia and returned to baseline levels by 14 dpi with the exception of GABA, Glutamate, total Creatine, and N-acetyl-aspartate. The injury induced impairment to cerebral energy metabolism indicated by a reduction in tCr was partially restored by exposure to hypobaria. Hypobaria combined with hyperoxia, however, reversed the improvements in cerebral energy metabolism gained with hypobaria alone and ultimately resulted in the lowest concentrations of both Glu and tCr measured during transport. In addition, only when hypobaria was combined with hyperoxia was there long-term reduction in tCr levels as well as a persistent reduction in CBF to the cortex and thalamus measured up to two weeks after the injury.

This may indicate the regulation of CBF in an integrative process that involves the marked influence of pulmonary gas exchange and cardiovascular function in addition to intracranial mediators of cerebral vessel resistance and flow. ⁵³ Some studies suggest that reduced CBF under hyperoxia is a form of autoregulation that provides neuroprotection by inhibiting O₂ delivery and ensuing oxidative stress. ⁵⁴ This scenario is not likely based on the finding that loss of neurochemicals were greatest in the groups that were exposed to hyperoxic hypobaria (Figure 6). Taken together, this may indicate that the region-specific acute changes observed in CBF may be due to hypobaria while delayed changes may reflect an adaptive response.

A distinct advantage of this study is the direct measurements of CBF and concentrations of specific cortical neurochemicals before, during, and after exposure to AERH. Additionally, while the outcome measures of many studies are restricted to one terminal endpoint after the initial TBI, our research longitudinally tracked changes in several parameters measured via MRI to demonstrate long term consequences of a single acute exposure to AERH. Limitations of this study include a relatively short observation period (14 days) after injury, precluding the evaluation of various long term injury outcomes. Given the growing evidence that females respond differently than males to traumatic brain injury, inflammation and altitude, a comparison of differences due to sex would strengthen future studies.^{55 –57} Although a strength of this study was the evaluation of a single homogenous trauma with one post injury transport, these findings may not be generalizable to TBI subjects with multiple injuries or those transported under varying conditions or timepoints following injury. Future studies involving the evaluation of multiple transports on military relevant compound traumas, including combinations of blast, brain impact, and hemorrhagic shock are warranted.

Conclusions and impact

This study provides innovative evidence that hypobaria exposure following TBI acts as a secondary insult that should be avoided if possible. Indiscriminate administration of 100% O₂ is not the answer, as is evident by the decrease in CBF and additional alterations of neurochemicals by exposure to the combination of hypobaria and hyperoxia. Our work indicates that O₂ should be titrated during hypobaric transport to a SpO2 level within the range of approximately 94 to 96% to avoid both hypoxia and hyperoxia, thereby minimizing secondary brain injury. ⁵⁸