Survival Rates and Biomarkers in a Large Animal Model of Traumatic Brain Injury Combined With Two Different Levels of Blood Loss : Shock

Secondary Logo

Journal Logo

Basic Science Aspects

Survival Rates and Biomarkers in a Large Animal Model of Traumatic Brain Injury Combined With Two Different Levels of Blood Loss

Mayer, Andrew R.∗,†,‡,§; Dodd, Andrew B.; Ling, Josef M.; Stephenson, David D.; Rannou-Latella, Julie G.; Vermillion, Meghan S.; Mehos, Carissa J.||; Johnson, Victoria E.; Gigliotti, Andrew P.; Dodd, Rebecca J.; Chaudry, Irshad H.∗∗; Meier, Timothy B.††,‡‡,§§; Smith, Douglas H.; Bragin, Denis E.∗,||||; Lai, Chen¶¶; Wagner, Chelsea L.¶¶; Guedes, Vivian A.¶¶; Gill, Jessica M.¶¶; Kinsler, Rachel∗∗∗

Author Information
SHOCK 55(4):p 554-562, April 2021. | DOI: 10.1097/SHK.0000000000001653



Traumatic brain injury (TBI) and severe blood loss resulting in hemorrhagic shock (HS) are leading causes of mortality and morbidity worldwide following trauma (1, 2). During concurrent TBI and HS (TBI+HS), a normally well-tolerated decrease in arterial pressure results in earlier and more severe reductions in cerebral blood flow, cerebral hypoxia, uncoupling of cerebral autoregulation, and a doubling of contusion volume as well as mortality rate (3, 4). Combined TBI+HS also interferes with cardiovascular compensatory mechanisms such as the modulation of vascular tone (5), increases respiratory abnormalities (6), and increases systematic inflammation (7).

Gross lesions as well as cellular and molecular changes are typical following TBI. Molecular changes include an increase in inflammatory markers and the release of glial fibrillary acidic protein (GFAP), neurofilament light chain (NFL), and ubiquitin C-terminal hydrolase (UCH-L1) into the bloodstream (8, 9). To date, large animal models of TBI+HS have primarily focused on contusional or mixed injury achieved with either controlled cortical impact (CCI) or fluid percussion injury (FPI), with all studies conducted exclusively in swine models (see previous review (10)). However, these primary contusion models fail to capture the diffuse injuries that typically occur following rapid acceleration/deceleration of the cranium and the resultant accumulation of mechanical shear/stress forces across the brain (11–13). Specifically, the use of craniotomy in most FPI and CCI studies results in prominent focal tissue destruction in the absence of hematoma and skull fracture, a constellation of findings infrequently observed in human trauma (12). Importantly, acceleration/deceleration models require species with higher brain mass to ensure that tissue loading forces approximate those experienced in humans (14). Even then, swine models of rotational head injury require an approximately 8-fold increase in acceleration to induce the same tissue deformations observed in the much larger human brain (15).

The lethal triad of hypothermia, acidosis, and coagulopathy represents the hallmark complications of HS (16), with the fundamental pathology for all HS states represented by a relative imbalance between systemic oxygen delivery and tissue oxygen consumption (i.e., oxygen debt (17, 18)). Approximately 75% of previous large animal TBI+HS studies utilized controlled blood loss models (typically withdrawal from femoral artery) to initiate HS, simulating Class III or IV trauma with an average of 35% to 45% of estimated total blood volume removed (reTBV) at an average rate of 2.18% reTBV/min (10). However, the average reported mortality rate (9.28%) in large animal TBI+HS models (10) during the untreated phase is much lower relative to the approximately 30% mortality rate reported for Class III and IV human trauma (17). The inflated survival rates in large animal models likely result from various experimental methodologies (e.g., utilization of mechanical ventilation, stopping and starting blood loss based on animal physiology such as mean arterial pressure [MAP] or heart rate [HR], administering small to medium fluid boluses) that are typically utilized experimentally but are not a component of the “untreated” phase of human trauma (10, 19). Other studies utilizing higher blood loss rates (50% or greater) in large animal models of HS alone have observed increased mortality rates (20).

The optimal resuscitation approaches and potential exacerbation of neural injury as a potential consequence of concurrent TBI+HS remain actively debated (16, 21). However, only a few studies have explicitly examined survivability across various levels of blood loss following TBI+HS (22–25), an important precursor to treatment studies. The primary objective of the current series of pilot studies was therefore to determine mortality rates following a novel TBI+HS model involving rapid head acceleration/deceleration and two different controlled blood loss levels (40% vs. 55%). As such, minimal treatment (e.g., no fluid resuscitation or mechanical ventilation) was provided both to more accurately determine true survival rates and to mimic care in more austere environments with prolonged evacuation times. Invasive hemodynamic measurements and blood-based biomarkers were also used to quantitatively document acute injury progression (9).


Animal preparation

All animal procedures (see Table 1 for timeline) were approved by the local Institutional Animal Care and Use Committee and the USAMRMC ORP Animal Care and Use Review Office prior to study initiation. Current results represent preliminary studies conducted as part of a larger study to examine the therapeutic effects of synthetic estrogen on TBI+HS and rough ground transport. Twenty-two sexually mature Yucatan swine (30.39 kg ± 2.25 kg; 11 males and 11 females) were fasted but provided ad libitum access to water for 6 to 12 h prior to surgical procedures. Animals were initially sedated with midazolam (0.5 mg/kg i.m. injection) and premedicated with buprenorphine (0.02 mg/kg i.m.). Animals were then intubated and maintained under general anesthesia for the duration of the experimental procedures (isoflurane: 5% induction, 1%–4% for maintenance combined with oxygen). Size 7 French catheters were placed into both femoral arteries, one for invasive blood pressure monitoring and the other for controlled blood loss and/or arterial blood sampling. To reduce clotting, each catheter was flushed approximately every 20 min with 1.5 mL of saline followed by 1.5 mL of heparinized saline (100 IU/mL). A single artificial breath was administered concurrently with the flush to fully inflate the lungs due to prolonged lateral recumbency.

Table 1 - Timeline of study tasks
Approximate time Tasks 40% 55% Sham
−90 min Surgery X X X
−15 min iSTAT X X X
0 min TBI X X
3 min Start BLP X X
25 min End BLP X X
35 min iSTAT X X X
85 min iSTAT X X X
145 min iSTAT X X X
205 min iSTAT X X X
295 min iSTAT X X
BD indicates blood draws for neural injury biomarkers; BLP, blood loss procedures; TBI, traumatic brain injury. The overall timeline was generally equivalent across cohorts until ∼205 min post-trauma when the single remaining TBI+55% animal was euthanized. Most TBI+40% (7/8) animals and all Sham animals were euthanized at ∼295 min post-trauma. Asterisks denote that data were not collected on all animals at that exact time due to non-survival, with the terminal sample being collected at either the final study time-point or euthanasia.

Animals were monitored for heart rate, end-tidal CO2, peripheral oxygen saturation, and body temperature with noninvasive sensors. Continuous invasive arterial pressure monitoring (IOX v2.9.5.73, emka Technologies, Paris, France) was used to capture heart rate (HR, a secondary outcome variable) and systolic and diastolic blood pressure, which were subsequently used to calculate primary (shock index [SI], pulse pressure [PP]), and secondary (MAP) hemodynamic outcome variables. Although the accuracy of the SI has been questioned in small animal models of TBI+HS (5), it is accurate for larger animals (26). Baseline invasive data were collected over a 5-min period approximately 15 min prior to the initiation of the polytrauma model.

Within the emka IOX platform, the invasive physiological data were smoothed and averaged over 30-s periods. All epochs were physically examined to determine an appropriate beat threshold (27) and to verify IOX had properly fit the blood pressure waveform. Average primary (SI, PP) and secondary outcome variables (MAP, HR) are reported for the last 30 s epoch of the baseline epoch prior to TBI, as soon as monitoring was re-established after TBI (coinciding with the start of the blood loss procedure), for the second phases of blood loss, and at approximately 1-h intervals after termination of the blood loss procedure to maximally correspond with blood draws (see Table 1 for actual times).

Blood samples were analyzed by a point-of-care device (i-STAT, CG8+ cartridges; Abbott Point of Care Inc, Princeton, NJ) and included potential hydrogen (pH), partial CO2 pressure (PCO2), bicarbonate (HCO3), sodium (Na), potassium (K), ionized calcium (iCa), glucose, and hematocrit, while lactate was measured with a separate device (Lactate Plus, Nova Biomedical Corp, Waltham, Mass). These biomarkers were collected at baseline, approximately 35 min following TBI (defined as trauma onset; see Table 1), approximately 85 min post-trauma, at 1-h intervals thereafter, and immediately prior to euthanasia based on individual animal survival times. In addition, plasma samples were collected at baseline, approximately 35 min post-trauma and immediately prior to euthanasia or upon animal death. Blood samples were centrifuged for 10 min at 2,500 relative centrifugal force, aliquoted into 400 μL to 500 μL tubes, and immediately stored in an −80°C freezer.

Protein concentrations in plasma samples were subsequently analyzed with a Simoa HD-1 Analyzer (Quanterix Simoa, Billerica, Mass) according to the manufacturer's protocol. A multiplex assay was used to analyze NFL, GFAP, UCH-L1, and total tau (Neuro 4-Plex A Kit, # 102153). Amyloid beta 40 and 42 (Aβ40 and Aβ42) were analyzed using single plex assays (Simoa Aβ40 2.0 Kit, #101672; Simoa Aβ42 2.0 Kit, #101664). Biomarker levels for each sample were measured in duplicate. Samples with a coefficient of variation higher than 30% in non-Sham animals post-trauma were excluded from subsequent analyses (see Supplemental Results,, whereas all Sham samples were utilized.

A closed-head TBI (28) was initiated via a pneumatic device (HYGE, Inc, Kittanning, Pa). Angular velocity for the TBI was acquired via an in-house data acquisition system using an ARS-06 angular rate sensor (Applied Technologies Associates, Albuquerque, NM) with a sampling frequency of 25 kHz that was rigidly mounted to the arm of the device. For the TBI exposure, all animals were placed in ventral recumbency (sternal) in sphynx position and maintained under isoflurane (1%–4%) with a midazolam bolus IV (0.1 mg/kg–0.5 mg/kg) as needed. Isoflurane was disconnected approximately 30 s prior to the TBI. The animals’ heads were secured to the linkage assembly of the HYGE device through a custom-made bite bar that converts linear to an angular (rotational) motion. The targeted angular peak velocity was 250 radians/s in the coronal plane for both the TBI+40% and TBI+55% cohorts (i.e., equivalent across groups) and was based on measurements from the rigidly mounted sensor. To reduce high-frequency noise, data were smoothed with a four-pole, Butterworth filter (channel frequency class = 1000 Hz) based on SAE-J211–1 recommendations (29) prior to identifying peak velocity. Time-to-peak (start time defined as 5% of peak velocity) and deceleration time (peak time to return to 5% of peak velocity) were identified as secondary variables.

Following the TBI exposure, animals were placed in lateral recumbency and subjected to arterial hemorrhage via controlled removal of approximately 55% (n = 8; 32.88 ± 1.18 kg; four females; hereafter referred to as TBI+55% cohort) or 40% (n = 8; 28.85 ± 1.51 kg; four females; hereafter referred to as TBI+40% cohort) of estimated total blood volume (eTBV = weight × 65 mL/kg). The controlled blood removal occurred via a peristaltic pump (Harvard Apparatus, Holliston, Mass) through either the left or right femoral arterial catheter. Blood loss was achieved in a biphasic fashion (50% of the total estimated amount [40% vs. 55%] removed in 7.5 min; remaining 50% in 15 min) to better approximate the normal physiological response during severe blood loss (30). Blood loss volume was subsequently confirmed by total blood weight (scalar factor = 1.06 g/mL). Animals were monitored for either 3 h (i.e., ∼205 min post-trauma; 55% model) or 4.5 h (∼295 min post-trauma; 40% model) following the blood loss procedure due to anticipated differences in survival rates. Animals were administered a combination of isoflurane (0%–4%) and midazolam (0.5 mg/kg/h to 1.5 mg/kg/h) immediately post-TBI, as well ketamine (8 mg/kg/h to 15 mg/kg/h) during the monitoring phase, to maintain a stable plane of anesthesia based on standard clinical monitoring signs.

A cohort of Sham animals (n = 6; 29.13 ± 0.70 kg; three females) underwent an identical series of procedures (e.g., catheter placement, placement on the HYGE device) with the exception of the actual TBI+HS trauma and were survived for ∼295 min post-Sham injury. Surviving animals from all cohorts were euthanized with intravenous Euthasol (1.0 mL/4.5 kg). With the exception of the use of U.S. Army-issued wool blankets to counteract hypothermia, no other supportive care (e.g., fluid resuscitation, mechanical ventilation, pausing of blood loss to maintain a certain MAP) was provided for any of the cohorts to better determine untreated survival times/rates in this novel TBI+HS model (10).

Tissue handling and neuropathological evaluation

All animals underwent necropsy and gross neuropathological findings were recorded. Tissue was processed to permit histological evaluation for the majority of animals. Specifically, retrograde aortic perfusion was performed with ambient temperature heparinized (10 IU/mL) phosphate-buffered saline, followed by ambient temperature 10% neutral buffered formalin (NBF). Following 24 h of fixation, the brain was removed from the skull and immersed in 10% NBF for 7 days. The brain was immediately trimmed (whole brain coronal blocks at 5 mm intervals), processed, and paraffin embedded.

Histopathology was performed on 8 μm thick sections, focusing at the level of the head of the caudate nucleus. Specifically, single immunohistochemistry labeling was performed to examine for extravasated serum proteins (immunoglobulin G [IgG]) as markers of blood–brain barrier integrity, axonal pathology (amyloid precursor protein [APP]), and microglia (ionized calcium-binding adaptor molecule 1 [IBA1]). Sections were first deparaffinized in xylene and rehydrated through descending grades of ethanol. Tissues were then quenched for endogenous peroxidase activity with 3% hydrogen peroxide for 15 min and heat-induced antigen retrieval was performed using Tris-EDTA buffer (pH 8.0) in a pressure cooker for 8 min. Tissues were blocked in 1% normal horse serum for 30 min at room temperature and incubated at 4°C overnight with a primary antibody recognizing amino acids 66 to 81 of the APP N-terminus (Millipore, Billerica, Mass; 1:90,000), microglial calcium binding protein IBA1 (Wako Chemicals; 1:2,500), or biotinylated anti-swine IgG (Vector Labs, Burlingame, Calif; 1:2,000). After washes with phosphate-buffered saline, sections were incubated with a biotinylated secondary antibody for 30 min, washed again in phosphate-buffered saline, and incubated with avidin–biotin complex (Universal Elite kit, Vector Labs, Burlingame, Calif) for 30 min. Sections were then visualized with 3, 3’-diaminobenzidine, counterstained with hematoxylin, dehydrated, and coverslipped. A negative control (no primary antibody) was included in each experiment to control for non-specific binding. Immunostained sections were imaged using an Olympus IX71 microscope.

Statistical plan

Generalized linear models (GLM) or linear mixed effects (LME) models were utilized for the majority of analyses with appropriate (Gaussian or Gamma) response distributions determined by the model fit. Cohorts (Sham vs. TBI+40% vs. TBI+55%) were first compared on key characteristics and baseline physiology to ensure effective matching (11). Survival rates up to ∼205 min post-trauma (Fig. 1) were assessed between the two TBI+HS cohorts using a Kaplan–Meier analysis with a Mantel-Cox log-rank significance test. Both primary and secondary physiological and neural biomarker data were compared with LME models using an unstructured covariance matrix to more accurately account for covariate dependent (Cohort) missingness (i.e., differences in survival rates). Neural biomarker data were log-transformed following the addition of a constant prior to analyses due to large differences in variance across injured and Sham cohorts. Change scores were calculated to better control for individual differences in baseline physiology.

Fig. 1:
This figure presents a survival curve (time in minutes) comparing TBI+40% (solid line) and TBI+55% (dashed line) cohorts post-trauma. The figure also includes the results from the Mantel-Cox test, indicating a significantly reduced survival rate in TBI+55% cohort at 205 min post-trauma, the last common time-point for both cohorts (i.e., TBI+40% cohort survived for a longer period). TBI indicates traumatic brain injury.

Baseline tests were not corrected for multiple comparisons to maintain a more liberal threshold during examination of nonexperimental variation. All other planned omnibus analyses were Bonferroni-corrected based on the number of comparisons performed for each set of biomarkers. Several variables contained extreme outliers (see Fig. 2 and Supplemental Figures, For these variables, analyses were repeated with the outliers removed and any discrepancies were noted in the manuscript. Critically, results for all primary measures were unaffected by outliers. Finally, additional generalized estimating equations (GEE) with either Gamma or Gaussian response distributions were performed within the Sham cohort to assess the variability of biomarkers as a function of anaesthesia and nontraumatic experimental procedures (e.g., surgery for catheter placement).

Fig. 2:
Panel (A) presents box-and-scatter plots for changes scores in plasma-based neural injury biomarker data (glial fibrillary acidic protein [GFAP], neurofilament light chain [NFL], ubiquitin carboxy-terminal hydrolase L1 [UCH-L1], and amyloid beta 40 and 42 [Aβ40 and Aβ42]) from the Sham (green diamonds) and TBI+40% (red diamonds) cohorts between baseline and ∼35 min (m) post-trauma samples (denoted with Δ), and between baseline and terminal samples (in all plotted animals, ∼295 min post-trauma). Asterisks (; Group main effect) and a deisis (; Group × Time interaction) are used to denote significant effects from Bonferroni-corrected (P ≤ 0.01) omnibus tests. Panel (B) displays immunohistochemistry results from the periventricular region at the level of the caudate nucleus for randomly selected animals from the Sham (n = 1) and TBI+40% (n = 2) cohorts. Data are presented for amyloid precursor protein (APP), Immunoglobulin G (IgG), and ionized calcium-binding adaptor molecule 1 (IBA1) antibodies. Similar immunohistochemical results were observed for animals in the TBI+55% cohort who survived for multiple hours postinjury (Supplemental Figure 2,


Demographics and baseline data

Results (Table 2) from uncorrected one factor GLMs indicated that the TBI+55% cohort weighed significantly more (Wald-χ2 = 70.91, P≤0.001) than both the Sham (P≤0.001) and TBI+40% (P≤0.001) cohorts, whereas Sham and TBI+40% did not differ significantly (P = 0.615). There were no statistical differences between the three groups in age, the amount of time animals were under anesthesia (e.g., catheter placement, experiment setup) prior to initialization of TBI, or hemodynamic (PP or SI) or blood-based physiological markers (all P's>0.05; Table 3). Similarly, there were no differences in neural biomarkers between the Sham and TBI+40% or the Sham and TBI+55% cohorts at baseline (all P's > 0.05).

Table 2 - Animal characteristics and HYGE parameters
Measure 40% (n = 8) 55% (n = 8) Sham (n = 6) P
Weight (kg) 28.85 ± 1.51 32.88 ± 1.18 29.13 ± 0.70 ≤0.001
Age (days) 181.88 ± 14.88 162.00 ± 26.68 188.67 ± 31.04 0.093
reTBV(%) 39.6 ± 1.4 54.7 ± 4.0 NA
 Decel Time (ms) 2.43 ± 0.11 3.38 ± 0.17 NA ≤0.001
 Velocity (rad/s) 254.56 ± 4.20 247.08 ± 6.73 NA 0.018
 Time to peak (ms) 6.13 ± 0.19 5.99 ± 0.08 NA 0.100
Decel Time indicates deceleration time; NA, not applicable; rad/s, radians per second; reTBV, removed estimated total blood volume. Note that P values are bolded if meeting Bonferroni correction among HYGE parameters (P ≤ 0.017).

Table 3 - Blood physiology and invasive arterial measurements
Baseline ∼25 min ∼85 min
Invasive 40% (n = 8) 55% (n = 8) Sham (n = 6) 40% (n = 8) 55% (n = 8) Sham (n = 6) 40% (n = 8) 55% (n = 4) Sham (n = 6) Omnibus P
PP 39.25 ± 5.70 44.00 ± 5.10 41.67 ± 7.55 9.25 ± 5.04 13.75 ± 10.00 41.17 ± 9.91 28.63 ± 9.44 25.00 ± 13.90 40.67 ± 9.73 0.003
SI 0.77 ± 0.26 0.91 ± 0.17 0.90 ± 0.38 4.35 ± 3.48 3.92 ± 2.25 0.62 ± 0.15 1.85 ± 0.69 3.51 ± 2.12 0.68 ± 0.28 0.004
HR (bpm) 87.75 ± 20.23 97.63 ± 21.10 89.17 ± 21.29 164.63 ± 36.02 153.75 ± 79.40 76.50 ± 7.53 166.38 ± 37.43 164.25 ± 79.31 78.17 ± 17.26 0.016
MAP (mm Hg) 95.00 ± 18.92 82.00 ± 12.12 81.33 ± 16.23 46.00 ± 21.62 27.38 ± 16.26 107.67 ± 23.62 80.12 ± 25.95 33.50 ± 21.75 100.00 ± 18.58 ≤0.001
Baseline ∼35 min ∼85 min
i-STAT 40% (n = 8) 55% (n = 8) Sham (n = 6) 40% (n = 8) 55% (n = 8) Sham (n = 6) 40% (n = 8) 55% (n = 4) Sham (n = 6) Omnibus P
pH 7.39 ± 0.04 7.40 ± 0.06 7.37 ± 0.04 7.33 ± 0.08 7.16 ± 0.18 7.42 ± 0.04 7.34 ± 0.10 7.31 ± 0.09 7.41 ± 0.03 0.003
PCO2 (mm Hg) 52.29 ± 5.05 53.71 ± 8.16 54.30 ± 4.64 37.74 ± 7.42 75.30 ± 42.87 53.40 ± 5.20 45.36 ± 4.59 30.23 ± 7.82 55.18 ± 4.43 0.001
HCO3 (mmol/L) 31.20 ± 2.36 33.14 ± 1.96 31.18 ± 2.93 19.86 ± 4.41 25.61 ± 7.91 34.43 ± 2.09 24.89 ± 4.54 15.73 ± 6.23 35.03 ± 1.90 ≤0.001
Na (mmol/L) 137.13 ± 2.17 138.00 ± 2.00 136.67 ± 1.51 133.38 ± 2.56 132.75 ± 2.49 136.67 ± 1.37 135.13 ± 2.85 134.50 ± 2.65 137.33 ± 1.37 ≤0.001
K (mmol/L) 4.06 ± 0.32 3.96 ± 0.39 4.08 ± 0.38 5.21 ± 0.64 6.33 ± 0.98 3.88 ± 0.23 4.38 ± 0.83 6.68 ± 1.80 3.82 ± 0.26 ≤0.001
iCa (mmol/L) 1.34 ± 0.06 1.38 ± 0.05 1.39 ± 0.04 1.21 ± 0.06 1.40 ± 0.11 1.36 ± 0.05 1.18 ± 0.06 1.08 ± 0.09 1.36 ± 0.04 ≤0.001
Glu (mg/dL) 96.00 ± 42.58 94.50 ± 41.10 73.50 ± 17.90 420.75 ± 66.49 393.50 ± 134.19 96.50 ± 6.35 250.25 ± 23.25 217.75 ± 45.81 91.50 ± 6.47 0.002
Hct (%PCV) 28.38 ± 1.77 27.63 ± 1.85 29.17 ± 1.33 26.88 ± 5.82 35.75 ± 5.92 28.50 ± 1.05 29.13 ± 4.97 32.50 ± 4.43 28.17 ± 0.75 0.010; 0.006
Lactate (mmol/L) 2.88 ± 1.80 2.06 ± 1.08 2.60 ± 0.67 8.63 ± 2.09 6.43 ± 1.94 1.28 ± 0.30 6.45 ± 1.96 10.40 ± 2.30 1.02 ± 0.41 ≤0.001
All times are approximate from the completion of the TBI, which is considered to be the onset of trauma (see Table 1).Glu indicates glucose; HCO3, bicarbonate; Hct, hermatocrit; HR, heart rate; iCa, ionized calcium; K, potassium; MAP, mean arterial pressure; Na, sodium; PCO2, partial pressure carbon dioxide; PP, pulse pressure; SI, shock index. = TBI+55% (n = 7). Values for omnibus P values are bolded if meeting Bonferroni correction among invasive blood (P ≤ 0.025) or iStat (P ≤ 0.00556) biomarkers. Symbols denote if the P value was derived from a main effect of Group () or Group × time interaction ().

The percentage (Fig. 1) of TBI+HS animals surviving to 205 min postinjury (i.e., duration of TBI+55% experiments) was significantly higher (χ2 = 13.02; P≤0.001) for the TBI+40% cohort (7/8; 87.5%) relative to TBI+55% cohort (1/8; 12.5%). Specifically, all swine in both cohorts survived the TBI and the 22.5 min controlled blood loss procedure. However, only four (50.0%) animals in the TBI+55% cohort survived up to 85 min post-trauma. The majority (five out of seven) of the TBI+55% animals and the single TBI+40% animal that expired prior to planned study endpoints were euthanized following extended apnea (i.e., more than 5 min). Statistical comparisons of TBI blood biomarker data were made separately for the TBI+40% (principal analyses) and TBI+55% cohorts due to potential confounding effects of shorter survival time/prolonged periods of hypoxia on neural markers (i.e., time-related confounds in protein accumulation). In contrast, comparisons of all non-TBI physiological biomarkers and invasive blood hemodynamic data were conducted across all three cohorts, but limited to 85 min post-trauma.

TBI-specific analyses

The TBI+40% cohort had significantly decreased average deceleration time (t14 = −13.05, P≤0.001) relative to the TBI+55% group. Peak velocity and time-to-peak were not significant following Bonferroni correction (0.05/3 = 0.017; Table 2; Supplemental Figure 1A, PP was significantly lower immediately following TBI (Wald-χ2 = 16.67, P≤0.001; Bonferroni correction (0.05/2 = 0.025) in the TBI+55% group relative to both Shams and TBI+40% (Supplemental Figure 1B, Results were not significant for SI (P = 0.76) immediately post-TBI. See Supplemental Results and Supplemental Figure 1C, for secondary hemodynamic measures of MAP (significant group differences when one outlier removed) and heart rate (no significant group effect).

Quantitative analyses of neural biomarkers (Supplemental Table 1,; all P's ≥ 0.01) were conducted using 2 × 2 LME models comparing TBI+40% to Shams or TBI+55% to Shams (Supplemental Materials, due to confounding effects of hypoxia/premature death (Bonferroni correction: 0.05/5 = 0.01). Results (Fig. 2A) indicated a significant interaction for NFL (F1,6.54 = 30.70; P = 0.001), characterized by statistically greater protein levels in TBI+40% relative to Shams at both time-points, but greater magnitude difference at 295 min post-trauma (P = 0.002) relative to 35 min post-trauma (P = 0.035). The TBI+40% cohort exhibited increased GFAP (F1,10.85 = 23.44; P = 0.001) and Aβ42 (F1,9.12 = 60.50; P≤0.001) relative to Shams across both time-points. Total tau was below detection levels for most animals at most time-points (data not analyzed).

The number of animals with gross brain pathology was 7/8 in the TBI+55% cohort and 8/8 in the TBI+40% cohort (Fisher exact P = 1.00). The most common abnormality was minimal to marked hemorrhage (medial to the dura matter) of the cerebellum, ventral brainstem, and/or lateral cerebral surface (15/16 animals). None of the animals in the Sham cohort exhibited gross pathological changes. Qualitative immunohistochemical results indicated axonal pathology, IgG extravasation, and microglia displaying morphological features of activation in TBI+40% (Fig. 2B) and TBI+55% (Supplemental Figure 2B, animals relative to Shams. Multifocal axonal pathology, characterized by accumulations of APP-positive axonal beads or varicosities, was observed throughout the section, with the largest accumulation of pathology surrounding the dorsolateral tip of the lateral ventricles. IgG extravasation was also commonly observed at the dorsolateral aspect of the lateral ventricles as well as in the depth of the sulci and subpial layers of the cortex of injured animals. Reactive microglia were visualized throughout the section of the TBI+40% and TBI+55% animals but were absent in Sham animals. Reactive microglia were primarily clustered around the radiation of the corpus callosum and gray/white matter interfaces.

Effects associated with severe blood loss

Results from the Bonferroni-corrected (0.05/2 = 0.025) primary invasive hemodynamic measurements (Table 3; Supplemental Figure 3A, indicated a significant Group × Time interaction for PP (F2,15.00 = 8.90, P = 0.003), with both TBI+40% (P≤0.001) and TBI+55% (P≤0.001) exhibiting decreased PP relative to the Sham cohort at 25 min post-trauma. At 85 min post-trauma, only the TBI+55% (P = 0.007) cohort demonstrated statistically lower PP values than Shams. In contrast, the TBI+40% and TBI+55% cohorts both exhibited significantly increased SI (F2,19.07 = 7.59, P = 0.004) from baseline relative to the Sham cohort across both the 25 and 85 min time points (all simple effects P's < 0.05). See Supplementary Materials, for analyses of secondary invasive hemodynamic variables (Supplemental Figure 1C,

Results (Bonferroni correction: 0.05/9 = 0.00556) indicated significant main effects (Table 3; Supplemental Figure 3B, for pH (F2,15.74 = 8.93, P = 0.003; Sham > TBI+40% > TBI+55%), Na (F2,10.44 = 29.18, P≤0.001; Sham > TBI+40% > TBI+55%), and K (F2,12.77 = 25.08, P≤0.001; Sham < TBI+40% < TBI+55%). Group × Time interactions were also present for PCO2 (F2,16.28 = 10.81, P = 0.001), HCO3 (F2,14.57 = 24.23, P≤0.001), iCa (F2,10.76 = 22.52, P≤0.001), glucose (F2,17.41 = 9.09, P = 0.002), and lactate (F2,15.39 = 23.05, P≤0.001).

In the first pattern (Table 3; Supplemental Figure 3C,, increased glucose and lactate as well as decreased HCO3 were observed for the TBI+55% and TBI+40% cohorts relative to Shams across both time points (all P's≤0.004). Glucose levels exhibited evidence toward recovery toward baseline for both groups at approximately 85 min post-trauma, whereas respective worsening for lactate (i.e., increasing; P = 0.001) and HCO3 (i.e., decreasing; P = 0.004) was observed for the TBI+55% relative to TBI+40% group. In the second pattern (Table 3; Supplemental Figure 3D,, TBI+40% displayed reduced iCa relative to both TBI+55% and Sham (all P's≤0.036) cohorts or reduced PCO2 relative to the TBI+55% cohort (P = 0.023) at 35 min post-trauma. However, a sharp reduction was subsequently observed for both iCa (all P's≤0.024) and PCO2 (all P's≤0.003) within the TBI+55% cohort relative to both cohorts at 85 min post-trauma. Neither main effects nor interactions were affected by the removal of outliers.

Effects of prolonged anesthesia and surgery

Unplanned GEE analyses (Bonferroni correction: 0.05/9 = 0.00556) examining time-dependent effects within the Sham cohort (e.g., surgery and/or prolonged anesthesia) indicated increased HCO3 (Wald-χ2 = 558.22; P≤0.001), Na (Wald-χ2 = 445.30; P≤0.001), and glucose (Wald-χ2 = 101.58; P≤0.001) relative to baseline, as well as decreased K (Wald-χ2 = 17.69; P = 0.001) and lactate (Wald-χ2= 277.53; P < 0.001). Finally, pH (Wald-χ2 = 28.82; P≤0.001) showed an initial decrease which then recovered to near baseline levels. GEE results for neural blood biomarkers (Bonferroni correction: 0.05/5 = 0.010; Supplemental Figure 4, indicated increased levels of NFL (Wald-χ2 = 37.29; P≤0.001) and GFAP (Wald-χ2 = 18.59; P≤0.001) across time relative to baseline, with decreased levels of UCH-L1 (Wald-χ2 = 14.84; P = 0.001) and Aβ42 (Wald-χ2 = 122.82; P≤0.001). Finally, Aβ40 (Wald-χ2 = 24.00; P≤0.001) demonstrated an increase at approximately 35 min post-trauma, while data collected at 295 min post-trauma were statistically equivalent to baseline (P = 0.529).


TBI and HS represent the two leading causes of death after injury in both civilian and military sectors (2). However, to date, only a handful of studies have explicitly examined survivability in large animal models of combined TBI+HS (22–25), and no models have incorporated the types of brain trauma (closed-head injury caused by dynamic acceleration) that typify human injury (10, 11, 31). Primary findings from the current study indicated large differences in survival rate between dynamic acceleration TBI+40% (12.5% mortality) versus TBI+55% (87.5% mortality) blood loss models. If replicated, current findings suggest that a TBI+55% model may be more appropriate for studies attempting to replicate severe trauma that requires immediate medical attention and standard fluid resuscitation protocols to promote survival (16), whereas the TBI+40% model may be better for scenarios in which advanced medical care and/or immediate access to resuscitation fluids are not available (e.g., prolonged field care (32)).

Current and previous results indicate that both systemic and cerebrovascular hemodynamic regulatory mechanisms are disrupted in TBI+HS, including decreased blood pressure regulation, the modulation of vascular tone, and reduced cardiovascular compensatory mechanisms (5, 33–35). In the current study, MAP was increased immediately following TBI in both injured cohorts relative to Shams, but then rapidly decreased at the start of the blood loss procedure. Measurements of pulse pressure and shock index were abnormal in both the TBI+40% and 55% cohorts immediately post-blood loss, with significant evidence of a dose-dependent response for pulse pressure at 85 min post-trauma. Similarly, a parametric response was observed for potassium, sodium, and pH across the 40% and 55% models, all of which are common blood-based biomarkers of shock, hypoxia, and cardiac functioning (16, 32, 36). Evidence of a worsening metabolic acidosis (i.e., increased lactate and decreased bicarbonate) was observed in the surviving members of the TBI+55% cohort at approximately 85 min, confirming the critical role for this process in mortality following severe blood loss (16, 19, 21). The majority of these remaining TBI+55% animals (3/4; 75%) expired within 3.5 h of the onset of polytraumatic injury, such that statistical power for all tests was limited even at relatively early postinjury time points.

Quantifying injury severity is challenging in animal models of isolated TBI (11, 12, 31) and is further compounded when considering subsequent prolonged periods of hemorrhagic shock and extended periods of anesthesia. However, the majority of animals in both the TBI+40% and TBI+55% cohorts had observable evidence of intracranial hemorrhage on gross necropsy, which would be classified as moderate to severe TBI based on most criteria for human TBI (37). The current study is the first to employ a dynamic acceleration TBI in combination with HS, historically considered to be the best model for generating diffuse axonal injury and modeling subsequent loss of consciousness (38). The increased mass and gyrencephalic nature of the swine brain also increases the likelihood that the current findings will translate to human trauma, including the aggregation of markers of neural pathology in both blood and brain tissue (10).

Specifically, similar to previous human (39) and swine (40) TBI studies, immunohistochemical and serum-based evidence of blood–brain barrier breach (IgG), diffuse axonal injury (APP accumulations/increased Aβ42) and inflammation (reactive microglia and increased GFAP) were present in the current combined TBI+HS models. Aβ is a 40 to 42 amino acid long peptide generated by successive cleavage of amyloid precursor protein by β-secretase followed by γ-secretase (41). TBI-induced vascular damage and the associated breakdown of the blood–brain barrier (42) may further stimulate cleavage of APP to release toxic species of Aβ, which occurred fairly rapidly (approximately 25–30 min) postinjury in the current study and have previously been associated with neural functioning post-TBI (43). Similarly, NFL serum levels have previously been shown to correlate with imaging markers of diffuse axonal injury in severe TBI (44) and remain elevated for prolonged periods in even milder forms of TBI (45). Although serum Aβ40 and UCH-L1 levels were not significantly elevated in the TBI+40% animals, this is more likely a result of low statistical power and high sample variance.

Immunohistochemical evidence (reactive microglia) of neuroinflammation was evident to a lesser extent in the current model due to shorter survival times (maximum time 295 min), with elevated plasma GFAP levels also increasing immediately postinjury. Reactive microglia have also been observed in the corpus callosum and surrounding gray-matter regions in postmortem human brains of individuals that expired during the chronic (28% of surveyed cases) injury phase (46), with most research indicating that microglia activity does not peak until several days postinjury (47). Finally, although preclinical rodent studies have indicated a rapid increase of oligomeric and phosphorylated tau aggregates as soon as 4 h after TBI relative to Shams (48), current findings indicated that total tau was below detection levels for the majority of animals sampled across all time-points. These discrepancies for total tau could be a result of the unknown cross-species homology of the total tau epitope in the Quanterix platform (see, or related to species-specific or model-specific differences.

Unexpectedly, all of the neural injury biomarkers exhibited small but significant time-dependent effects in the Sham cohort, most likely as a result of prolonged periods of anesthesia and/or surgical procedures. Prolonged periods of anesthesia are necessary for the ethical conductance of large animal research, but introduce confound not present in human trauma scenarios (10). Preclinical data suggest that inhaled anesthetics disrupt brain vascular endothelial cells, increase blood–brain barrier permeability and enhance amyloid beta oligomerization rates (49–51), which have been further associated with postoperative cognitive decline. A human study suggested that inhaled isoflurane increased Aβ40 in cerebral spinal fluid whereas desflurane decreases Aβ42 post-surgery (52). Similar changes in point-of-care markers were also observed for several HS biomarkers (pH, HCO3, Na, K, glucose, and lactate), which can be more readily attributed to periods of prolonged metabolic depression and inactivity. It is important to note that only six animals were included in the Sham cohort, rendering these findings preliminary in nature and in need of future replication given the variability observed in some of the biomarker data. However, these results highlight the need for equally powered control groups in large animal studies to ensure that biomarkers are either injury-specific or of sufficient magnitude to exceed the levels produced by non-specific experimental effects of anesthesia and Sham procedures.

The majority of animals in both TBI+HS cohorts met moribund criteria due to prolonged periods of apnea rather than from cardiac arrest. Respiratory depression is the most common cause of death in moderate-to-severe TBI preclinical models (31), with a recent study indicating that respiratory complications are also much more frequent following TBI+HS relative to HS alone in humans (6). It has been known for decades that hyperventilation protects against mortality during hypovolemia through mitigation of acidosis (53). However, agonal gasping is also present at the end-stages of HS, and may increase cerebral blood flow through reduced intrathoracic pressure (54). Unlike all previous large animal studies of TBI+HS (10), the current study did not utilize mechanical ventilation or other supportive care (maintenance of certain MAP and/or HR levels during blood removal and/or small bolus infusions) to more accurately estimate true survival rates. Mechanical ventilation likely interferes with the typical pathophysiological cascade of TBI+HS both by overriding the compensatory hyperventilation response that is critical to restoring systemic acid-base balance (55) and potentially mitigating the respiratory depression that occurs in TBI with high rotational components (31). Importantly, anesthesia and mechanical ventilation are also related, as mechanical ventilation is frequently used to counter the known respiratory depressive properties of most anesthetic agents in preclinical studies. The current protocol partially mitigated this confound through utilization of anesthetic agents with reduced effects on respiratory depression (i.e., midazolam and ketamine) postinjury rather than isoflurane.

The current preliminary studies were limited by several factors. Foremost, although the current model was purposely developed to more closely mirror typical human trauma scenarios, it still involved intubation, prolonged anesthesia, and small doses of heparin to maintain catheter patency, none of which are present in most pretreatment scenarios in human care. Second, animals were not randomized to experimental cohorts, and all data for the TBI+55% cohort were collected first based on a priori assumptions about survival rates from HS only studies (20). Importantly, experimental methodologies for injury and immediate post injury monitoring (i.e., up to 205 min post-trauma) were similar across cohorts, minimizing any potential confounds in the current study. However, this ordering effect resulted in unanticipated differences for some animal characteristics (i.e., weight), as well as secondary TBI injury parameters (i.e., deceleration time). While the former was likely a result of litter effects from the ordering of the experiment, it is important to note that animals were of the same age and, therefore, similar sexual maturity. The latter was likely as a result of routine machine maintenance (replacement of parts) that occurred between cohorts, but likely did not influence survival rates given that the deceleration time was longer in the TBI+55% cohort relative to the TBI+40% cohort. Finally, differences in survival rates between the TBI+40% versus TBI+55% likely affected neural biomarkers immunohistochemistry results due to the dynamic changes in protein accumulations that occur as a function of time postinjury during the acute phase of TBI (9).

In summary, understanding the complex pathophysiology and best treatment options for patients with combined TBI+HS remains a critical need (2). Current results provide preliminary evidence for different survival rates in 40% versus 55% controlled blood loss models when coupled with a realistic, closed-head, dynamic acceleration model, and minimal supportive care in a large animal model. Thus, 55% or higher blood loss rates commonly utilized in previous hemorrhage only models (20) may not be tenable for future intervention studies mimicking prolonged care following TBI+HS. Findings also highlight the role of respiratory crisis in the context of multiple pathologies following concurrent TBI+HS, which may have been artificially mitigated through the prevalent use of mechanical ventilation in previous studies (10). Thus, multiple therapeutic strategies ultimately may be necessary to combat the different pathophysiological effects that TBI and HS individually introduce (16, 21).


The authors also thank Mandy Housand, Fawn Reed, Sheila Alonzo, Jaime Leo, and Jimmy Gonzales for help with data collection.


1. Eastridge BJ, Mabry RL, Seguin P, Cantrell J, Tops T, Uribe P, Mallett O, Zubko T, Oetjen-Gerdes L, Rasmussen TE, et al. Death on the battlefield (2001–2011): implications for the future of combat casualty care. J Trauma Acute Care Surg 73:S431–S437, 2012.
2. Shafi S, Collinsworth AW, Richter KM, Alam HB, Becker LB, Bullock MR, Ecklund JM, Gallagher J, Gandhi R, Haut ER, et al. Bundles of care for resuscitation from hemorrhagic shock and severe brain injury in trauma patients-Translating knowledge into practice. J Trauma Acute Care Surg 81:780–794, 2016.
3. Giri BK, Krishnappa IK, Bryan RM Jr, Robertson C, Watson J. Regional cerebral blood flow after cortical impact injury complicated by a secondary insult in rats. Stroke 31:961–967, 2000.
4. Matsushita Y, Bramlett HM, Kuluz JW, Alonso O, Dietrich WD. Delayed hemorrhagic hypotension exacerbates the hemodynamic and histopathologic consequences of traumatic brain injury in rats. J Cereb Blood Flow Metab 21:847–856, 2001.
5. McMahon CG, Kenny R, Bennett K, Little R, Kirkman E. Effect of acute traumatic brain injury on baroreflex function. Shock 35:53–58, 2011.
6. Galvagno SM, Fox EE, Appana SN, Baraniuk S, Bosarge PL, Bulger EM, Callcut RA, Cotton BA, Goodman M, Inaba K, et al. Outcomes after concomitant traumatic brain injury and hemorrhagic shock: a secondary analysis from the pragmatic, randomized optimal platelets and plasma ratios trial. J Trauma Acute Care Surg 83:668–674, 2017.
7. Liu H, Xiao X, Sun C, Sun D, Li Y, Yang M. Systemic inflammation and multiple organ injury in traumatic hemorrhagic shock. Front Biosci (Landmark Ed) 20:927–933, 2015.
8. Blennow K, Hardy J, Zetterberg H. The neuropathology and neurobiology of traumatic brain injury. Neuron 76:886–899, 2012.
9. Zetterberg H, Blennow K. Fluid biomarkers for mild traumatic brain injury and related conditions. Nat Rev Neurol 12:563–574, 2016.
10. Mayer AR, Dodd AB, Vermillion MS, Stephenson DD, Chaudry IH, Bragin DE, Gigliotti AP, Dodd RJ, Wasserott BC, Shukla P, et al. A systematic review of large animal models of combined traumatic brain injury and hemorrhagic shock. Neurosci Biobehav Rev 104:160–177, 2019.
11. Smith DH, Hicks RR, Johnson VE, Bergstrom DA, Cummings DM, Noble LJ, Hovda D, Whalen M, Ahlers ST, LaPlaca M, et al. Pre-clinical traumatic brain injury common data elements: Toward a common language across laboratories. J Neurotrauma 32:1725–1735, 2015.
12. Vink R. Large animal models of traumatic brain injury. J Neurosci Res 96:527–535, 2018.
13. Ghajari M, Hellyer PJ, Sharp DJ. Computational modelling of traumatic brain injury predicts the location of chronic traumatic encephalopathy pathology. Brain 140:333–343, 2017.
14. Johnson VE, Meaney DF, Cullen DK, Smith DH. Animal models of traumatic brain injury. Handb Clin Neurol 127:115–128, 2015.
15. Browne KD, Chen XH, Meaney DF, Smith DH. Mild traumatic brain injury and diffuse axonal injury in swine. J Neurotrauma 28:1747–1755, 2011.
16. Cannon JW. Hemorrhagic Shock. N Engl J Med 378:370–379, 2018.
17. American College of Surgeons Committee on Trauma. Advanced Trauma Life Support Program for Doctors. 7th ed.Chicago, IL: American College of Surgeons; 2004.
18. Cairns CB. Rude unhinging of the machinery of life: metabolic approaches to hemorrhagic shock. Curr Opin Crit Care 7:437–443, 2001.
19. Hildebrand F, Andruszkow H, Huber-Lang M, Pape HC, van Griensven M. Combined hemorrhage/trauma models in pigs-current state and future perspectives. Shock 40:247–273, 2013.
20. Miller M, Keith J, Berman J, Burlington DB, Grudzinskas C, Hubbard W, Peck C, Scott C, Chaudry IH. Efficacy of 17alpha-ethynylestradiol-3-sulfate for severe hemorrhage in minipigs in the absence of fluid resuscitation. J Trauma Acute Care Surg 76:1409–1416, 2014.
21. Fulop A, Turoczi Z, Garbaisz D, Harsanyi L, Szijarto A. Experimental models of hemorrhagic shock: a review. Eur Surg Res 50:57–70, 2013.
22. Stern SA, Zink BJ, Mertz M, Wang X, Dronen SC. Effect of initially limited resuscitation in a combined model of fluid-percussion brain injury and severe uncontrolled hemorrhagic shock. J Neurosurg 93:305–314, 2000.
23. Stern S, Rice J, Philbin N, McGwin G, Arnaud F, Johnson T, Flournoy WS, Ahlers S, Pearce LB, McCarron R, et al. Resuscitation with the hemoglobin-based oxygen carrier, HBOC-201, in a swine model of severe uncontrolled hemorrhage and traumatic brain injury. Shock 31:64–79, 2009.
24. Zink BJ, Stern SA, Wang X, Chudnofsky CC. Effects of ethanol in an experimental model of combined traumatic brain injury and hemorrhagic shock. Acad Emerg Med 5:9–17, 1998.
25. Zink BJ, Sheinberg MA, Wang X, Mertz M, Stern SA, Betz AL. Acute ethanol intoxication in a model of traumatic brain injury with hemorrhagic shock: effects on early physiological response. J Neurosurg 89:983–990, 1998.
26. Frohlich M, Driessen A, Bohmer A, Nienaber U, Igressa A, Probst C, Bouillon B, Maegele M, Mutschler M. Is the shock index based classification of hypovolemic shock applicable in multiple injured patients with severe traumatic brain injury?-an analysis of the TraumaRegister DGU((R)). Scand J Trauma Resusc Emerg Med 24:148, 2016.
27. Smith DA, Craige ER. Mechanism of the dicrotic pulse. Heart 56:531–534, 1986.
28. Cullen DK, Harris JP, Browne KD, Wolf JA, Duda JE, Meaney DF, Margulies SS, Smith DH. A porcine model of traumatic brain injury via head rotational acceleration. Methods Mol Biol 1462:289–324, 2016.
29. Society of Automotive Engineers: SAE J211-1 (1995): Instrumentation for Impact Test, Part 1, Electronic Instrumentation, 1995. Society of Automotive Engineers, Inc.
30. Frankel DA, Acosta JA, Anjaria DJ, Porcides RD, Wolf PL, Coimbra R, Hoyt DB. Physiologic response to hemorrhagic shock depends on rate and means of hemorrhage. J Surg Res 143:276–280, 2007.
31. Xiong Y, Mahmood A, Chopp M. Animal models of traumatic brain injury. Nat Rev Neurosci 14:128–142, 2013.
32. Baker BL, Powell D, Riesberg J, Keenan S. Prolonged field care working group fluid therapy recommendations. J Spec Oper Med 16:112–117, 2016.
33. Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the autonomic and cardiovascular systems in acute brain injury. Am J Physiol 275:R1287–R1292, 1998.
34. Yuan XQ, Wade CE. Influences of traumatic brain injury on the outcomes of delayed and repeated hemorrhages. Circ Shock 35:231–236, 1991.
35. Yuan XQ, Wade CE, Clifford CB. Suppression by traumatic brain injury of spontaneous hemodynamic recovery from hemorrhagic shock in rats. J Neurosurg 75:408–414, 1991.
36. Rocha Filho JA, Nani RS, D’Albuquerque LA, Holms CA, Rocha JP, Sa Malbouisson LM, Machado MC, Carmona MJ, Auler Junior JO. Hyperkalemia accompanies hemorrhagic shock and correlates with mortality. Clinics (Sao Paulo) 64:591–597, 2009.
37. Mayer AR, Quinn DK, Master CL. The spectrum of mild traumatic brain injury: a review. Neurology 89:623–632, 2017.
38. Gennarelli TA, Thibault LE, Adams JH, Graham DI, Thompson CJ, Marcincin RP. Diffuse axonal injury and traumatic coma in the primate. Ann Neurol 12:564–574, 1982.
39. Johnson VE, Stewart W, Smith DH. Traumatic brain injury and amyloid-beta pathology: a link to Alzheimer's disease? Nat Rev Neurosci 11:361–370, 2010.
40. Smith DH, Chen XH, Nonaka M, Trojanowski JQ, Lee VM, Saatman KE, Leoni MJ, Xu BN, Wolf JA, Meaney DF. Accumulation of amyloid beta and tau and the formation of neurofilament inclusions following diffuse brain injury in the pig. J Neuropathol Exp Neurol 58:982–992, 1999.
41. Karran E, De SB. The amyloid cascade hypothesis: are we poised for success or failure? J Neurochem 139: (suppl 2): 237–252, 2016.
42. Wojnarowicz MW, Fisher AM, Minaeva O, Goldstein LE. Considerations for experimental animal models of concussion, traumatic brain injury, and chronic traumatic encephalopathy-these matters matter. Front Neurol 8:240, 2017.
43. Brody DL, Magnoni S, Schwetye KE, Spinner ML, Esparza TJ, Stocchetti N, Zipfel GJ, Holtzman DM. Amyloid-beta dynamics correlate with neurological status in the injured human brain. Psychobiology 321:1221–1224, 2008.
44. Ljungqvist J, Zetterberg H, Mitsis M, Blennow K, Skoglund T. Serum neurofilament light protein as a marker for diffuse axonal injury: results from a case series study. J Neurotrauma 34:1124–1127, 2017.
45. Shahim P, Zetterberg H, Tegner Y, Blennow K. Serum neurofilament light as a biomarker for mild traumatic brain injury in contact sports. Neurology 88:1788–1794, 2017.
46. Johnson VE, Stewart JE, Begbie FD, Trojanowski JQ, Smith DH, Stewart W. Inflammation and white matter degeneration persist for years after a single traumatic brain injury. Brain 136:28–42, 2013.
47. Donat CK, Scott G, Gentleman SM, Sastre M. Microglial activation in traumatic brain injury. Front Aging Neurosci 9:208, 2017.
48. Hawkins BE, Krishnamurthy S, Castillo-Carranza DL, Sengupta U, Prough DS, Jackson GR, DeWitt DS, Kayed R. Rapid accumulation of endogenous tau oligomers in a rat model of traumatic brain injury: possible link between traumatic brain injury and sporadic tauopathies. J Biol Chem 288:17042–17050, 2013.
49. Acharya NK, Goldwaser EL, Forsberg MM, Godsey GA, Johnson CA, Sarkar A, DeMarshall C, Kosciuk MC, Dash JM, Hale CP, et al. Sevoflurane and Isoflurane induce structural changes in brain vascular endothelial cells and increase blood-brain barrier permeability: Possible link to postoperative delirium and cognitive decline. Brain Res 1620:29–41, 2015.
50. Nemoto EM, Stezoski SW, MacMurdo D. Glucose transport across the rat blood-brain barrier during anesthesia. Anesthesiology 49:170–176, 1978.
51. Eckenhoff RG, Johansson JS, Wei H, Carnini A, Kang B, Wei W, Pidikiti R, Keller JM, Eckenhoff MF. Inhaled anesthetic enhancement of amyloid-beta oligomerization and cytotoxicity. Anesthesiology 101:703–709, 2004.
52. Zhang B, Tian M, Zheng H, Zhen Y, Yue Y, Li T, Li S, Marcantonio ER, Xie Z. Effects of anesthetic isoflurane and desflurane on human cerebrospinal fluid Abeta and tau level. Anesthesiology 119:52–60, 2013.
53. Brandfonbrener M, Whang R. Effect of respiratory alkalosis on survival in hemorrhagic shock. Circ Res 21:461–468, 1967.
54. Suzuki M, Funabiki T, Hori S, Aikawa N. Spontaneous gasping increases cerebral blood flow during untreated fatal hemorrhagic shock. Resuscitation 80:109–112, 2009.
55. Gutierrez G, Das A, Ballarino G, Beyzaei-Arani A, Turkan H, Wulf-Gutierrez M, Rider K, Kaya H, Amdur R. Decreased respiratory rate variability during mechanical ventilation is associated with increased mortality. Intensive Care Med 39:1359–1367, 2013.

Biomarkers; hypovolemia; large animal models; mortality; shock; traumatic brain injury

Supplemental Digital Content

Copyright © 2020 by the Shock Society