Advertisement

Serum neurofilament light as a biomarker of vulnerability to a second mild traumatic brain injury

Open AccessPublished:November 16, 2022DOI:https://doi.org/10.1016/j.trsl.2022.11.008

      Abstract

      A second mild traumatic brain injury (mTBI) sustained prior to neuropathological recovery can lead to exacerbated effects. Without objective indicators of this neuropathology, individuals may return to activities at risk of mTBI when their brain is still vulnerable. With axonal injury recognised as a neuropathological hallmark of mTBI, we hypothesized that serum levels of neurofilament light (NfL), a highly sensitive biomarker of axonal injury, may be predictive of vulnerability to worse outcomes in the event of a second mTBI. Given this hypothesis is difficult to test clinically, we used a two-hit model of mTBI in rats and staggered inter-injury intervals by 1-, 3-, 7- or 14-days. Repeat-mTBI rats were dichotomised into NfLhigh (NfL>median at the time of re-injury) and NfLlow (NfL<median) groups, with behavior and NfL levels analysed throughout the 28-days, followed by ex vivo diffusion tensor imaging. NfL levels at the time of the second mTBI were found to be predictive of vulnerability to re-injury, with NfLhigh rats displaying more neurological signs and a greater potentiation of NfL levels after the second mTBI. Importantly, this potentiation phenomenon remained even when limiting analyses to rats with longer inter-injury intervals, providing evidence that vulnerability to re-injury may not be exclusively dependent on inter-injury interval. Finally, NfL levels correlated with, and were predictive of, the severity of neurological signs following the second mTBI. These findings provide evidence that measurement of NfL during mTBI recovery may be reflective of the vulnerability to a second mTBI, and as such may have utility to assist return to sport, duty and work decisions.

      Keywords

      Introduction

      Mild traumatic brain injury (mTBI) is a highly prevalent injury in collision sports and in the military [

      Faul M, Wald MM, Xu L, Coronado VG. Traumatic brain injury in the United States; emergency department visits, hospitalizations, and deaths, 2002-2006. 2010.

      ,
      • Meehan 3rd, WP
      • Mannix R.
      Pediatric concussions in United States emergency departments in the years 2002 to 2006.
      ,
      • Levin HS
      • Diaz-Arrastia RR.
      Diagnosis, prognosis, and clinical management of mild traumatic brain injury.
      ,
      • Hoge CW
      • McGurk D
      • Thomas JL
      • Cox AL
      • Engel CC
      • Castro CA.
      Mild traumatic brain injury in U.S. Soldiers returning from Iraq.
      ]. There is mounting evidence that a history of mTBI is a major risk factor for suffering a future mTBI, and for experiencing worse symptoms for a longer duration should another mTBI occur [
      • Dretsch MN
      • Silverberg ND
      • Iverson GL.
      Multiple Past Concussions Are Associated with Ongoing Post-Concussive Symptoms but Not Cognitive Impairment in Active-Duty Army Soldiers.
      ,
      • Miller KJ
      • Ivins BJ
      • Schwab KA.
      Self-reported mild TBI and postconcussive symptoms in a peacetime active duty military population: effect of multiple TBI history versus single mild TBI.
      ,
      • Iverson GL
      • Gardner AJ
      • Terry DP
      • Ponsford JL
      • Sills AK
      • Broshek DK
      • et al.
      Predictors of clinical recovery from concussion: a systematic review.
      ]. The latency between mTBIs appears to be critical, with evidence suggesting that the impact of a previous mTBI on symptom severity and duration lessens as time elapses [
      • Silverberg ND
      • Lange RT
      • Millis SR
      • Rose A
      • Hopp G
      • Leach S
      • et al.
      Post-concussion symptom reporting after multiple mild traumatic brain injuries.
      ,
      • Eisenberg MA
      • Andrea J
      • Meehan W
      • Mannix R
      Time Interval Between Concussions and Symptom Duration.
      ]. The risk of mTBIs in short succession creating a significant cumulative burden has led to the hypothesis that the pathophysiology of mTBI may create a ‘window of increased cerebral vulnerability’ to repeated injury [
      • Silverberg ND
      • Lange RT
      • Millis SR
      • Rose A
      • Hopp G
      • Leach S
      • et al.
      Post-concussion symptom reporting after multiple mild traumatic brain injuries.
      ,
      • Eisenberg MA
      • Andrea J
      • Meehan W
      • Mannix R
      Time Interval Between Concussions and Symptom Duration.
      ,
      • Kamins J
      • Bigler E
      • Covassin T
      • Henry L
      • Kemp S
      • Leddy JJ
      • et al.
      What is the physiological time to recovery after concussion? A systematic review.
      ,
      • Vagnozzi R
      • Tavazzi B
      • Signoretti S
      • Amorini AM
      • Belli A
      • Cimatti M
      • et al.
      Temporal window of metabolic brain vulnerability to concussions: mitochondrial-related impairment–part I.
      ,
      • Giza CC
      • Prins ML
      • Hovda DA.
      It's Not All Fun and Games: Sports, Concussions, and Neuroscience.
      ,
      • Vagnozzi R
      • Signoretti S
      • Tavazzi B
      • Floris R
      • Ludovici A
      • Marziali S
      • et al.
      Temporal window of metabolic brain vulnerability to concussion: a pilot 1H-magnetic resonance spectroscopic study in concussed athletes–part III.
      ]. As such, there is a widely recognized need to identify objective tools that can help detect and monitor the heterogeneous neuropathology of mTBI to assist individual clinical decisions, such as when it is safe to return to play, duty or any other at-risk activity after concussion [
      • McCrory P
      • Meeuwisse W
      • Dvorak J
      • Aubry M
      • Bailes J
      • Broglio S
      • et al.
      Consensus statement on concussion in sport-the 5(th) international conference on concussion in sport held in Berlin, October 2016.
      ].
      Diffuse axonal injury is now recognized as a key neuropathological feature of mTBI [
      • VE Johnson
      • Stewart W
      • Smith DH.
      Axonal pathology in traumatic brain injury.
      ,
      • Hill CS
      • Coleman MP
      • Menon DK.
      Traumatic Axonal Injury: Mechanisms and Translational Opportunities.
      ,
      • Miller DR
      • Hayes JP
      • Lafleche G
      • Salat DH
      • Verfaellie M.
      White matter abnormalities are associated with chronic postconcussion symptoms in blast-related mild traumatic brain injury.
      ]. We and others have recently shown that serum levels of neurofilament light (NfL), a highly sensitive and specific biomarker of neuroaxonal damage, are elevated after mTBI and remain elevated beyond typical symptom resolution [
      • McDonald SJ
      • O'Brien WT
      • Symons GF
      • Chen Z
      • Bain J
      • Major BP
      • et al.
      Prolonged elevation of serum neurofilament light after concussion in male Australian football players.
      ,
      • Shahim P
      • Tegner Y
      • Marklund N
      • Blennow K
      • Zetterberg H.
      Neurofilament light and tau as blood biomarkers for sports-related concussion.
      ,
      • Zetterberg H
      • Blennow K.
      Fluid biomarkers for mild traumatic brain injury and related conditions.
      ,
      • Shahim P
      • Zetterberg H
      • Tegner Y
      • Blennow K.
      Serum neurofilament light as a biomarker for mild traumatic brain injury in contact sports.
      ,
      • Clarke GJB
      • Skandsen T
      • Zetterberg H
      • Einarsen CE
      • Feyling C
      • Follestad T
      • et al.
      One-Year Prospective Study of Plasma Biomarkers From CNS in Patients With Mild Traumatic Brain Injury.
      ,
      • McDonald SJ
      • Piantella S
      • O'Brien WT
      • Hale MW
      • O'Halloran P
      • Kinsella G
      • et al.
      Clinical and blood biomarker trajectories after concussion: New insights from a longitudinal pilot study of professional flat-track jockeys.
      ]. While consistetly elevated at a group level, the extent and trajectory of NfL changes after mTBI appears to be variable between individuals [
      • McDonald SJ
      • O'Brien WT
      • Symons GF
      • Chen Z
      • Bain J
      • Major BP
      • et al.
      Prolonged elevation of serum neurofilament light after concussion in male Australian football players.
      ,
      • Shahim P
      • Tegner Y
      • Marklund N
      • Blennow K
      • Zetterberg H.
      Neurofilament light and tau as blood biomarkers for sports-related concussion.
      ,
      • McDonald SJ
      • Piantella S
      • O'Brien WT
      • Hale MW
      • O'Halloran P
      • Kinsella G
      • et al.
      Clinical and blood biomarker trajectories after concussion: New insights from a longitudinal pilot study of professional flat-track jockeys.
      ]. We hypothesised that the extent and duration of this neuropathology after mTBI may be a mechanism of a variable vulnerability to repeat mTBI (rmTBI), and as such, serum NfL quantification may have utility to not only identify axonal damage in individuals with suspected mTBI, but also to track neuropathological progression, recovery, and vunerability to further injury.
      This cerebral vulnerability hypothesis is however very difficult to investigate in humans. As such, here we utilised a two-mTBI model in rats, and investigated whether serum NfL levels at the time of a second impact were related to neuropathological and behavioral outcomes. We hypothesised that outcomes would be significantly worse in rats with higher levels of serum NfL at the time of the second mTBI.

      Materials and Methods

      Animals

      Eighty-four male adolescent Sprague-Dawley rats were housed in cages of three on a 12-hour light/dark cycle. Food and water were available ad libitum. All procedures were approved by the Alfred Medical Research and Educational Precinct Animal Ethics Committee (E/2081/2021/M). Animal studies are reported in compliance with the ARRIVE guidelines and the Australian code of practice for the care and use of animals for scientific purposes by the Australian National Health and Medical Research Council.

      Experimental groups

      Rats were randomly allocated to sustain two mTBI procedures (rmTBI), one sham and one mTBI (single mTBI), or two sham procedures (sham). The rmTBI rats were further broken down into four groups defined by their inter-injury interval: 1-, 3-, 7-, or 14-days. To mitigate the effect of age on injury severity, all groups contained two sub-groups; the first group sustained their first mTBI/sham procedure at the same age, and the second group sustained their second mTBI/sham procedure at the same age. This produced a consistent mean PND (i.e. PND 47) of each mTBI/sham procedure.
      Six rats that recorded a score of zero for the video analysis of mTBI-related signs (described below) as well as deemed by an investigator at the time of mTBI procedure to sustain an unsuccessful mTBI were excluded from analyses. One rat was excluded from all analyses due to death immediately following the mTBI procedure, and another due to a cage-associated injury. As such, 76 rats were included in the final analyses: rmTBI-1d (n=13), rmTBI-3d (n=12), rmTBI-7d (n=13), rmTBI-14d (n=13), single-mTBI (n=12) and sham (n=13). To assess the primary question of the association between serum NfL levels at the time of injury and subsequent outcomes, all rmTBI groups (irrespective of inter-injury group) were dichotomised by their NfL level at the time of the second mTBI; the first of which had serum NfL levels less than the median (NfLlow), and those that had greater than the median (NfLhigh). With six 0-day blood samples unable to be collected, final group sizes for the primary analysis were n=23 rats for NfLhigh and NfLlow groups. An overview of the study protocol for the primary analysis of NfLhigh and NfLlow comparisons is shown in Figure 1.
      Figure 1:
      Figure 1Overview of study protocol.
      Created with BioRender.com

      Awake closed head injury (ACHI) model of mTBI

      The ACHI model is a well characterized and clinically relevant model of concussive-like injury [
      • O'Brien WT
      • Pham L
      • Brady RD
      • Bain J
      • Yamakawa GR
      • Sun M
      • et al.
      Temporal profile and utility of serum neurofilament light in a rat model of mild traumatic brain injury.
      ,
      • Pham L
      • Shultz SR
      • Kim HA
      • Brady RD
      • Wortman RC
      • Genders SG
      • et al.
      Mild Closed-Head Injury in Conscious Rats Causes Transient Neurobehavioral and Glial Disturbances: A Novel Experimental Model of Concussion.
      ,
      • Pham L
      • Wright DK
      • O'Brien WT
      • Bain J
      • Huang C
      • Sun M
      • et al.
      Behavioral, axonal, and proteomic alterations following repeated mild traumatic brain injury: Novel insights using a clinically relevant rat model.
      ]. It avoids the confounds of anaesthetic and craniotomy present in other models, features clinically relevant biomechanics and induces behavioral deficits that are typically transient. The ACHI impact/sham protocol, and two day habituation protocol was performed as previously described [
      • Pham L
      • Shultz SR
      • Kim HA
      • Brady RD
      • Wortman RC
      • Genders SG
      • et al.
      Mild Closed-Head Injury in Conscious Rats Causes Transient Neurobehavioral and Glial Disturbances: A Novel Experimental Model of Concussion.
      ]. All behavioural testing described herein was performed at consistent time-points relative to the final sham/mTBI.

      Video quantification of observable neurological signs of mTBI

      By avoiding anaesthetic, the ACHI model of mTBI allows for clinically relevant neurological signs to be observed and quantified immediately after the injury. As such, we developed a scale for quantifying neurological signs in rats that are observable on video. This scale was developed following the ‘International consensus definitions of video signs of concussion’ [
      • Davis GA
      • Makdissi M
      • Bloomfield P
      • Clifton P
      • Echemendia RJ
      • Falvey EC
      • et al.
      International consensus definitions of video signs of concussion in professional sports.
      ]. This scale quantifies four mTBI-related signs: i) post-impact seizure/tonic posturing; ii) motionlessness; iii) forelimb incoordination whereby one, or both forelimbs were unable to grasp a 2cm beam); and iv) hind limb incoordination whereby the hind limbs slip off the 2cm beam and/or an inability to balance on the beam. The presence of each sign was given a score of one, with a maximum score of four. All videos were analysed by an investigator blinded to the experimental conditions.

      Rotarod

      Sensorimotor function was assessed using the rotarod task [
      • Ah Kim H
      • Semple BD
      • Dill LK
      • Pham L
      • Dworkin S
      • Zhang SR
      • et al.
      Systemic treatment with human amnion epithelial cells after experimental traumatic brain injury.
      ]. Briefly, rats were placed on a rod accelerating from 4 rpm to 40 rpm over a five-minute period, with latency to fall the primary outcome. Three trials were performed for each session conducted at baseline, and at 1-, 2-, 6-, 13-, and 27-days post the second ACHI/sham.

      Water maze

      Spatial memory was assessed using a water maze previously described [
      • Pham L
      • Wright DK
      • O'Brien WT
      • Bain J
      • Huang C
      • Sun M
      • et al.
      Behavioral, axonal, and proteomic alterations following repeated mild traumatic brain injury: Novel insights using a clinically relevant rat model.
      ]. Briefly, a 175cm diameter pool filled with tap water (28 degrees Celsius) with four different visual stimuli suspended above the edge of the pool, and a hidden platform located 3.5 cm below the water surface. Rats were placed at pseudo-random starting locations for ten trials per day (acquisition testing on day 4, reversal on day 5), with the visual stimuli designed to be used as cues to locate the hidden platform. TopScanLite version 2.0 was used to track rats, with latency to find the platform used a measure of spatial memory.

      Elevated plus maze (EPM)

      The EPM is a measure of anxiety-like behavior and was performed 12-days following the second mTBI/sham following protocols previously described [
      • Pham L
      • Shultz SR
      • Kim HA
      • Brady RD
      • Wortman RC
      • Genders SG
      • et al.
      Mild Closed-Head Injury in Conscious Rats Causes Transient Neurobehavioral and Glial Disturbances: A Novel Experimental Model of Concussion.
      ,
      • Pham L
      • Wright DK
      • O'Brien WT
      • Bain J
      • Huang C
      • Sun M
      • et al.
      Behavioral, axonal, and proteomic alterations following repeated mild traumatic brain injury: Novel insights using a clinically relevant rat model.
      ]. Briefly, a ‘plus-shaped’ maze featuring two 30cm high walls (closed arms) and two arms without walls (open arms) was used, with reduced time spent in the open arms an indicator of anxiety-like behavior.

      Blood and tissue collection

      Whole blood was collected from the lateral tail vein 1-hour prior to the second mTBI/sham (i.e. 0-day), and 3-, 7-, and 14-days afterwards. Rats were lightly anaesthetised via isoflurane inhalation, a 23” gauge needle was then inserted into the lateral tail vein, and blood slowly collected into BD SST™ microtainer tubes. Blood was centrifuged at 1,500g for 10 minutes, with serum collected and stored at -80 degrees Celsius. At 28-days rats were euthanized as described previously [
      • Pham L
      • Wright DK
      • O'Brien WT
      • Bain J
      • Huang C
      • Sun M
      • et al.
      Behavioral, axonal, and proteomic alterations following repeated mild traumatic brain injury: Novel insights using a clinically relevant rat model.
      ], with whole blood collected via cardiac puncture and brains fixed in paraformaldehyde.

      Magnetic Resonance Imaging (MRI)

      DTI was performed with a 9.4T Bruker MRI using a 2-shot echo planar imaging sequence. Diffusion-weighting was performed in 61 directions with δ = 4.2 ms, Δ = 12 ms and b-values = 2000 and 4000 s/mm2. Two b = 0 (b0) volumes were also acquired. Other imaging parameters were adjusted to give an isotropic resolution of 250 μm: repetition time = 8 s; echo time = 50 ms; field of view = 3.2×2.4 cm2; and 48 axial slices. A subsequent b0 image was acquired with the same imaging parameters and the phase reversed for distortion correction using FSL's topup [
      • Andersson JL
      • Skare S
      • Ashburner J.
      How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging.
      ]. Image processing was performed as previously described using FSL and MRtrix3 software [
      • Tournier JD
      • Smith R
      • Raffelt D
      • Tabbara R
      • Dhollander T
      • Pietsch M
      • et al.
      MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation.
      ,
      • Zamani A
      • O'Brien TJ
      • Kershaw J
      • Johnston LA
      • Semple BD
      • Wright DK.
      White matter changes following experimental pediatric traumatic brain injury: an advanced diffusion-weighted imaging investigation.
      ]. The mean fractional anisotropy (FA)-value was determined for the ipsilateral and contralateral genu, body and splenium of the corpus callosum.

      Serum NfL quantification

      NfL was quantified on a SIMOA® HD-X Analyzer™ using NfL Advantage kits (Quanterix, Billerica, MA, USA) following manufacturer's instructions. All samples were above the lower limit of detection (0.038 pg/mL). Two samples were run across all plates with an average inter-plate coefficient of variation (CV) of 6.2%, and the intraplate CV of 4.8%.

      Statistics

      To investigate differences between the NfLlow and NfLhigh group on behavioral outcomes a Mann-Whitney U-test was used with an α-value of 0.05 for signs of mTBI and EPM outcomes, and an adjusted α-value of 0.0083 for rotarod and 0.01 for water maze. To assess the relationship between NfL levels pre- and post- the second mTBI, NfL concentrations underwent a natural logarithmic transformation (Ln) and levels across time were assessed with a two-way ANOVA. Post-hoc analyses were performed with Bonferroni's multiple comparisons. The magnitude of serum NfL change (ΔNfL) was assessed with a Mann-Whitney U-test. MRI outcomes for the three regions of the corpus callosum (genu, body and splenium) were analysed on hemispheres independently with a paired t-test with an adjusted α-value of 0.017.
      Spearman's correlation assessed for the association between observable neurological signs and serum NfL levels. The number of neurological signs in the rmTBI rats was also dichotomized as mild (≤2 signs, n=18), or severe (>2 signs, n=28), with classification ability of serum NfL assessed using an area under the receiver operating characteristic (AUROC) analyses and closest-to-(0,1) corner approach used to determine the optimal cut-off points [
      • Rota M
      • Antolini L
      • Valsecchi MG.
      Optimal cut-point definition in biomarkers: the case of censored failure time outcome.
      ].
      To assess the association between inter-injury interval on behavioral and MRI outcomes, a Kruskal-Wallis H test was used to assess for group differences for all outcomes with an α-value of 0.05 for video signs of mTBI and EPM outcomes, and an adjusted α-value of 0.0083 for the rotarod, 0.01 for water maze and 0.017 for MRI outcomes. To assess the relationship between inter-injury interval and subsequent NfL outcomes, NfL concentrations were Ln transformed and a mixed-effects model with the Giesser-Greenhouse correction was performed. Post-hoc analyses were performed with Bonferroni's multiple comparisons test. An overall group effect of injury on ΔNfL was assessed using a Kruskal-Wallis H test. MRI outcomes were assessed with a Mann-Whitney U-test. All statistical analyses were performed with GraphPad Prism GraphPad Prism version 8.0.2 for Windows (GraphPad Software, CA).

      Results

      NfLhigh rats had more observable neurological signs and a potentiated serum NfL profile after a second mTBI

      The median serum NfL levels of all rmTBI rats at the time of the second mTBI (i.e., day-0) was found to be 32.0 pg/mL. Rats were then split into two groups (Figure 2A); the first of which had serum NfL levels less than 32.0 pg/mL (NfLlow), and those that had greater than 32.0 pg/mL (NfLhigh).
      Figure 2:
      Figure 2NfL levels at the time of a second mTBI are indicative of the resultant mTBI severity.
      (A) Box plots of serum NfL levels at the time of injury for each of the four rmTBI inter-injury groups. The four rmTBI inter-injury groups were collapsed to create an all rmTBI group. The median NfL level for all rmTBI rats was 32.0 pg/mL. Rats with NfL levels lower than the group median were allocated to the NfLlow group, and rats with levels greater than the group median were allocated to the NfLhigh group. (B) Group level plots (median, IQR) show the temporal profile of NfL at the time of second injury, and at post second-mTBI time-points of 3-, 7-, 14-, and 28-days for NfLhigh, NfLlow, and sham rats. NfLhigh rats had higher NfL levels at post second mTBI time-points of 3-, 7-, and 14-days (all p<0.0001) compared with NfLlow rats. (C) Violin plots show that the magnitude of change in NfL from the time of injury to 3-days post the second mTBI (ΔNfL) was greater in the NfLhigh group than NfLlow group (p=0.0448). (D) ΔNfL was also greater in the NfLhigh group than NfLlow group when analysis was restricted to 7- and 14-day interval rats only (p=0.01). (E) A greater number of observable neurological signs for the second mTBI was found in the NfLhigh group when compared with the NfLlow group (p=0.0044). * p<0.05, ** p<0.01, **** p<0.0001.
      Dichotomised by NfL levels at the time of injury, there was an overall effect of group (F(1,44)=44.09, p<0.0001), time (F(2.69,118)=195, p<0.0001) and group X time interaction (F(4,176)=20.8, p<0.0001) on the post second-mTBI profile of Ln serum NfL (Figure 2B). Post-hoc analyses revealed that serum Ln NfL levels were elevated in the NfLhigh group at day-3 (95%CI: 1.70-0.663, p<0.0001), day-7 (95%CI: 1.30-0.459, p<0.0001), day-14 (95%CI: 1.04-0.361, p<0.0001) but not day-28 (95%CI:0.562-0.116, p=0.415) compared with the NfLlow group.
      To assess the magnitude of serum NfL change after a second mTBI, levels at the time of the second mTBI (i.e., day-0) were subtracted from the serum NfL levels at day-3 (ΔNfL = NfL3d – NfL0d). NfLhigh rats were found to have a greater ΔNfL than NfLlow rats (U=173, Median NfLlow=13.5 pg/mL, Median NfLhigh=44.3 pg/mL, p=0.0448; Figure 2C). The magnitude of ΔNfL was further assessed excluding the shorter inter-injury interval rmTBI groups (i.e., 1-day and 3-day interval), with the rationale being that: i) 7-day and 14-day intervals may be more comparable to typical rmTBI exposure in humans (although days in rats have been compared to weeks in humans), and ii) our previous data with a single injury revealed that NfL levels have plateaued at 7- and 14-days, and therefore ΔNfL values are less impacted by the dynamic changes that occur more acutely. For the rmTBI 7-day and 14-day rats only, the time of second mTBI NfL median was found to be 22.6 pg/mL, with rats subsequently grouped as NfLhigh+ (>22.6 pg/mL, n=12) or NfLlow+ (<22.6 pg/mL, n=12). NfLhigh+ rats were found to have a greater ΔNfL than NfLlow+ rats (U=28, Median NfLlow*=21.3 pg/mL, Median NfLhigh*=46.1 pg/mL, p=0.01; Figure 2D).
      Finally, NfLhigh rats had a greater number of observable neurological signs of mTBI immediately after the second mTBI when compared to NfLlow rats (U=143.5, Median NfLlow = 2, median NfLhigh = 3, p=0.0044; Figure 2E).

      Serum NfL levels at the time of a second mTBI correlate with, and are prognostic of, observable neurological signs

      Spearman's correlations analyses tested the association between the number of observable neurological signs, and serum NfL levels at the time of second mTBI (day-0), and at 3-, 7-, 14-, and 28-days post-injury. For all rmTBI rats, neurological signs correlated with NfL at the time of injury (Figure 3A; r=0.500, p=0.0004), and at 3- (Figure 3B; r=0.636, p<0.0001), 7- (Figure 3C; r=0.522, p=0.0002) and 14- (Figure 3D; r=0.498, p=0.0004), but not 28-days (Figure 3E; r=0.273, p=0.0662).
      Figure 3:
      Figure 3The relationship between serum NfL levels and mTBI-related observable neurological signs.
      The number of observable neurological signs immediately after the second mTBI correlated with serum NfL levels at the time of second mTBI (A), and post second-mTBI time-points of 3- (B), 7- (C) and 14-days (D). The number of observable neurological signs immediately after the second mTBI did not correlate to NfL levels at 28-days (E). A regression line is added to each scatterplot to assist visualisation of the spread of data. (F) Serum NfL at the time of injury had a good prognostic ability for the severity of visual signs (i.e., ≤2 versus >2 signs) immediately after the second mTBI with an AUROC value of 0.73 (sensitivity = 0.64; specificity = 0.78) at a 33.3 pg/mL cut-off.
      The ability of time of injury NfL levels to predict injury severity, as determined by observable neurological signs immediately after the second mTBI (dichotomised by ≤2 or >2 signs) was assessed with the AUROC analysis (Figure 3F). The area under the curve was 0.73 (95% CI: 0.59-0.87, p=0.009) with a sensitivity of 0.64 (95%CI : 0.46-0.79) and specificity of 0.78 (95% CI :0.55-0.91) at the optimal cut-off of 33.3 pg/mL.

      Serum NfL levels at the time of a second mTBI versus sub-acute behavior and chronic white matter integrity

      To assess the relationship between NfL level at the time of injury with behavioral outcomes in the sub-acute stages after the second mTBI, comparisons were made between NfLlow and NfLhigh groups for each task. No differences were found in water maze latency to escape for the acquisition (day 4) or reversal testing (day 5). No differences were found for the time spent in the EPM open arm (day 12), nor for rotarod latency to fall at baseline, or 1-, 2-, 6-, 13-, or 27-days post the second mTBI. Sub-acute behavioral data are detailed in Supplementary Figure 2 and Supplementary Table 1.
      We investigated brain white matter integrity at 28-days using DTI. Region of interest was the three regions of the corpus callosum per hemisphere (genu, body and splenium; Figure 4). Despite trends of reduced FA in NfLhigh rats, with an α-value of 0.017 (0.05 / 3 regions), no significant differences in FA were found on the left (i.e. ipsilateral side to injury) genu (p=0.0458), body (p=0.269) or splenium (p=0.0450) of the corpus callosum. Similarly, no significant differences were found on the right hemisphere for the genu (p=0.0706), body (p=0.846) or splenium (p=0.293).
      Figure 4:
      Figure 4Chronic white matter integrity in rats with low versus high serum NfL levels at the time of a second mTBI
      In the left hemisphere, although not surviving multiple comparison corrections (α=0.017), a trend towards a decrease in FA for the NfLhigh group was found in the genu (A; p=0.046), and splenium (C; p=0.045), but not the body of the corpus callosum (B; p=0.269). In the right hemisphere, a trend for a decrease in FA in the NfLhigh group was found in the genu (D; p=0.071), but not the body (E; p=0.846), or splenium of the corpus callosum (F; p=0.293).

      The relationship between inter-injury interval and serum NfL outcomes

      For the effect of inter-injury interval on serum NfL levels across time (Figure 5A), there was an overall effect of time (F(2.48,170)=195, p<0.0001), group (F(5,70)=14.5, p<0.0001) and time X group interaction (F(20,274)=17.2, p<0.0001). Notably, multiple comparisons revealed that all groups were different to sham at 3- and 7-days. At 14-days, while NfL levels in the rmTBI groups remained higher than sham group, the single mTBI group was no longer different, and at 28-days, only the rmTBI (1-day interval) group was different to sham. See Supplementary Table 2 for a full summary of post-hoc results. No effect of inter-injury interval was found for the ΔNfL ((H(3)=6.77, P=0.0795; Figure 5B).
      Figure 5:
      Figure 5The relationship between inter-injury interval and serum NfL outcomes
      (A) Box plots show the distribution of serum NfL levels at the time of second mTBI, and post second-mTBI time-points of 3-, 7-, 14-, and 28-days for sham, single mTBI, and the four inter-injury interval rmTBI groups. For full details of the post-hoc results for each time-point see supplementary table 2. (B) No inter-injury interval effect was found for ΔNfL.

      Inter-injury interval versus sub-acute behavior and chronic white matter integrity

      No differences were found between inter-injury interval groups for the number of observable neurological signs immediately after the second mTBI, EPM time in open arm, latency to find the hidden platform in the water maze, and latency to fall from the rotarod (Supplementary Figure 3 and Supplementary Table 3). For the effect of inter-injury interval on FA in the corpus collosum, no group differences were found (Supplementary Figure 4).

      Discussion

      Following an mTBI, the brain can be highly vulnerable to a subsequent mTBI, with this increased vulnerability postulated to be due to pathophysiological changes induced by the first mTBI [
      • Silverberg ND
      • Lange RT
      • Millis SR
      • Rose A
      • Hopp G
      • Leach S
      • et al.
      Post-concussion symptom reporting after multiple mild traumatic brain injuries.
      ,
      • Eisenberg MA
      • Andrea J
      • Meehan W
      • Mannix R
      Time Interval Between Concussions and Symptom Duration.
      ,
      • Kamins J
      • Bigler E
      • Covassin T
      • Henry L
      • Kemp S
      • Leddy JJ
      • et al.
      What is the physiological time to recovery after concussion? A systematic review.
      ,
      • Vagnozzi R
      • Tavazzi B
      • Signoretti S
      • Amorini AM
      • Belli A
      • Cimatti M
      • et al.
      Temporal window of metabolic brain vulnerability to concussions: mitochondrial-related impairment–part I.
      ,
      • Giza CC
      • Prins ML
      • Hovda DA.
      It's Not All Fun and Games: Sports, Concussions, and Neuroscience.
      ,
      • Vagnozzi R
      • Signoretti S
      • Tavazzi B
      • Floris R
      • Ludovici A
      • Marziali S
      • et al.
      Temporal window of metabolic brain vulnerability to concussion: a pilot 1H-magnetic resonance spectroscopic study in concussed athletes–part III.
      ]. As such, there is a need for objective biomarkers that can detect the severity of injury and monitor pathophysiological recovery following an mTBI, to ultimately assist decisions surrounding the safe return to sport, duty or other activities at-risk of mTBI [
      • McCrory P
      • Meeuwisse W
      • Dvorak J
      • Aubry M
      • Bailes J
      • Broglio S
      • et al.
      Consensus statement on concussion in sport-the 5(th) international conference on concussion in sport held in Berlin, October 2016.
      ]. Here, we quantified serum NfL, a clinically applicable circulating biomarker of axonal injury, to determine whether levels of this protein at the time of a second mTBI are associated with worse outcomes following rmTBI. Overall, our findings supported the hypothesis that high serum NfL levels at the time of a second mTBI are associated with a greater vulnerability to axonal damage and functional impairments. We found that NfLhigh rats displayed more observable neurological signs of mTBI immediately after the second impact, and that serum NfL levels at the time of re-injury had a good prognostic ability for the severity of these signs. Moreover, we found that the relative magnitude of NfL increase from time of re-injury to 3-days (ΔNfL) was higher in NfLhigh rats. Serum NfL levels also correlated with observable neurological signs of injury, and the elevation of serum NfL in rats that had sustained two mTBIs at a shorter inter-injury interval (i.e., 1d) was relatively prolonged.

      Serum NfL levels at the time of a second impact are related to resultant mTBI severity

      We assessed the magnitude of change of serum NfL levels from the time of re-injury to 3-days (ΔNfL) to provide insights into the extent of axonal damage resulting from the second mTBI alone. We found that the ΔNfL values were greater in NfLhigh rats, indicating that these rats sustained a greater severity of injury from the second impact. It is important to recognise that serum NfL levels at the time of the second impact were not stagnant, with previous and current rat ACHI data indicating a peak of serum NfL at 1-day followed by an exponential decline through to 14-days [
      • O'Brien WT
      • Pham L
      • Brady RD
      • Bain J
      • Yamakawa GR
      • Sun M
      • et al.
      Temporal profile and utility of serum neurofilament light in a rat model of mild traumatic brain injury.
      ]. It is therefore likely that the rate of decline in NfL was greatest in rats with the shorter inter-injury intervals (i.e., 1- and 3-day intervals). As the NfLhigh group largely consisted of 1- and 3-day interval rmTBI rats (i.e., 16/23 rats), the finding that the NfLhigh rats had a greater ΔNfL value, is despite the likely faster clearance of NfL levels at the time of the second mTBI. Furthermore, when removing the 1- and 3-day interval rats from the ΔNfL analysis, the group differences were still present. As such, this finding provides evidence that NfL levels at the time of second impact may be predictive of the extent of the resultant axonal damage.
      It is important to recognize that serum NfL levels resulting from the first mTBI are not only an indicator of vulnerability to re-injury but also likely reflective of the severity of the initial injury. The variability of serum NfL responses after the first mTBI is likely attributed to some combination of minor inconsistencies in the injury procedure itself and individual factors related to cerebral vulnerability. Such factors may include subtle differences in anatomy (e.g., skull thickness, brain size, neck musculature), physiology (e.g., basal inflammation) and environment (e.g., stressors and socialization) between rats. Combinations of these factors may contribute to a more severe first mTBI, or more prolonged axonal pathology, in some rats when compared to others, with this severity spectrum reflected in serum NfL levels. After the second mTBI, these same factors also likely contribute to outcome, but in this case, our findings indicate that the extent of axonal pathology, as reflected by serum NfL levels at the time of the second mTBI, are key to vulnerability to a subsequent impact. Importantly, the aforementioned ΔNfL findings are suggestive of a greater degree of injury due to the second mTBI alone in NfLhigh rats. As such, we conclude that serum NfL levels after a single mTBI are likely reflective of injury severity, but also a vulnerability to re-injury that is related to the extent and duration of axonal pathology induced by the initial mTBI. To further understand the contribution of initial injury severity to vulnerability, future research may consider a protocol in which rats are given one of two impact settings (e.g., one 20% greater than the other) for the first mTBI, followed by the same settings for the second mTBI. In addition, studies may consider implementing both serial measures of NfL prior to the second mTBI and longer inter-injury intervals to further understand the association between initial injury severity and ongoing axonal pathology on the duration and potential cessation of axonal vulnerability.
      It has been previously shown that observable neurological signs of mTBI (i.e., tonic posturing, motionlessness, motor incoordination, etc.) correlate well to clinical outcomes such as orientation, concentration, and recall [
      • Reyes J
      • Mitra B
      • Makdissi M
      • Clifton P
      • Nguyen JVK
      • Harcourt P
      • et al.
      Visible Signs of Concussion and Cognitive Screening in Community Sports.
      ]. We found that observable neurological signs at the time of the second mTBI where more substantial in NfLhigh when compared with NfLlow rats. Removing the group dichotomy from analysis, pre-injury NfL levels also correlated well with neurological signs after the second impact. The lack of lasting behavioral deficits in this study limits conclusions on the ability of NfL to predict deficits beyond the acute stages. This finding should however be considered in context, with behavioral testing in rodents often lacking sensitivity to behavioral impairments than can be reported or detected in clinical mTBI [
      • Shultz SR
      • McDonald SJ
      • Corrigan F
      • Semple BD
      • Salberg S
      • Zamani A
      • et al.
      Clinical Relevance of Behavior Testing in Animal Models of Traumatic Brain Injury.
      ,
      • DeWitt DS
      • Hawkins BE
      • Dixon CE
      • Kochanek PM
      • Armstead W
      • Bass CR
      • et al.
      Pre-Clinical Testing of Therapies for Traumatic Brain Injury.
      ]. It is possible that alternative tasks or testing at different time-points may have revealed longer lasting deficits. Nonetheless, given that video signs of injury alone have shown some predictive value in clinical mTBI, our finding that serum NfL levels were predictive of the number of neurological signs at the time second mTBI in rats provides preliminary evidence of a potential prognostic utility of this biomarker.
      Our DTI analysis focused on FA of the corpus callosum, due to previous ACHI model studies finding FA differences in this region [
      • Pham L
      • Wright DK
      • O'Brien WT
      • Bain J
      • Huang C
      • Sun M
      • et al.
      Behavioral, axonal, and proteomic alterations following repeated mild traumatic brain injury: Novel insights using a clinically relevant rat model.
      ]. Although not surviving multiple comparison corrections, 3/6 corpus callosum regions had trends towards lower FA in the NfLhigh rats. It is plausible that the findings of a possible decrease in FA is indicative of a longer-lasting central disturbance to white matter integrity in rats with high NfL levels at the time of injury; however, this trend requires further investigation.

      The effect of inter-injury interval on the profile of serum NfL

      A secondary aim was to investigate the effect of timing between mTBIs alone on subsequent behavioral, structural, and NfL outcomes. While it is well recognised that timing is an important factor to rmTBI outcomes [
      • Silverberg ND
      • Lange RT
      • Millis SR
      • Rose A
      • Hopp G
      • Leach S
      • et al.
      Post-concussion symptom reporting after multiple mild traumatic brain injuries.
      ,
      • Eisenberg MA
      • Andrea J
      • Meehan W
      • Mannix R
      Time Interval Between Concussions and Symptom Duration.
      ], and multiple studies have demonstrated heightened cumulative effects when mTBIs are separated by 1 to 3 days [
      • Vagnozzi R
      • Tavazzi B
      • Signoretti S
      • Amorini AM
      • Belli A
      • Cimatti M
      • et al.
      Temporal window of metabolic brain vulnerability to concussions: mitochondrial-related impairment–part I.
      ,
      • Vagnozzi R
      • Signoretti S
      • Tavazzi B
      • Floris R
      • Ludovici A
      • Marziali S
      • et al.
      Temporal window of metabolic brain vulnerability to concussion: a pilot 1H-magnetic resonance spectroscopic study in concussed athletes–part III.
      ,
      • Weil ZM
      • Gaier KR
      • Karelina K.
      Injury timing alters metabolic, inflammatory and functional outcomes following repeated mild traumatic brain injury.
      ,
      • Prins ML
      • Alexander D
      • Giza CC
      • Hovda DA.
      Repeated mild traumatic brain injury: mechanisms of cerebral vulnerability.
      ,
      • Longhi L
      • Saatman KE
      • Fujimoto S
      • Raghupathi R
      • Meaney DF
      • Davis J
      • et al.
      Temporal window of vulnerability to repetitive experimental concussive brain injury.
      ,
      • Grant DA
      • Serpa R
      • Moattari CR
      • Brown A
      • Greco T
      • Prins ML
      • et al.
      Repeat Mild Traumatic Brain Injury in Adolescent Rats Increases Subsequent beta-Amyloid Pathogenesis.
      ], the exact period in which the brain remains susceptible to a second mTBI is not known. The finding that rmTBI with the shortest inter-injury interval (1-day) had the most prolonged profile of serum NfL, may support the hypothesis that the brain is at an increased vulnerability to rmTBI with a shorter inter-injury interval. However, a limitation of this component of the study is the within-group variability in the injury severity. Due to the potential for a discrepancy in age (and hence weight) between the first and last injury (14-days), groups were split in half so that one half of all rats sustained their first mTBI/sham at the same age, and for the second half all rats sustained their second mTBI/sham at the same age. This meant that all groups had a very similar mean age for the two mTBIs. While this allows for most appropriate group level comparisons, it did however, likely contribute to a greater variability within interval groups and may account for the lack of statistical differences between interval groups.
      Our primary analysis of NfLhigh versus NfLlow rats does however shed some light on the duration of vulnerability. Firstly, when removing the 1- and 3-day interval rats from the ΔNfL analysis, the NfLhigh versus NfLlow group differences were still present in 7- and 14-day interval rats, indicating a degree of axonal vulnerability that may exist beyond the commonly reported 3-day window in rats. Moreover, in the full data set analysis (i.e., all rmTBI irrespective of inter-injury) interval, it is notable that NfLhigh rats were not exclusively derived from those given a shorter recovery period (i.e., 7/23 NfLhigh rats had recovery periods of 7- or 14-days). These findings provide some evidence that axonal vulnerability in the event of a second mTBI is unlikely to be solely dependent on time since the initial mTBI, and that monitoring serum NfL levels during recovery at an individual level may be necessary to inform decisions such as when it is safe return to play decisions after sport-related concussion.

      Limitations

      It is important to consider the limitations of the study. Firstly, separating rmTBI rats by the median NfL level was an a priori hypothesis of the study; it is however, an arbitrary threshold and may not represent the most appropriate threshold to predict vulnerability to re-injury. Nonetheless, this median NfL level (32.0 pg/mL) was very similar to the optimal cut-off concentration in AUROC analysis (i.e., 33.3 pg/mL). While serum NfL is a widely recognised and highly sensitive biomarker of axonal injury, this study did not directly assess central measures of axonal damage at the time of injury. In addition, due to differences in group sizes as well as a high within group variability for the inter-injury groups, it is not possible to determine whether monitoring timing or serum NfL levels may have more utility to mitigate the cumulative effects of rmTBI. Notably, although serum NfL reflects the extent of axonal pathology, it is not possible to determine if this or another accompanying and potentially correlated pathology, such as glial, metabolic, vascular or inflammatory alterations, is key to cerebral vulnerability. Related to this, it is important to note that we have not compared the utility of NfL to other promising biomarkers of mTBI, such as GFAP and S100B, and that literature to date indicates that a panel of markers that reflect different aspects of pathology and have different kinetics, are likely to have greatest utility in the diagnosis and management of mTBI. Finally, this study investigated young male rodents only; future studies are required to determine the influence of age and biological sex.

      Conclusion

      Overall, this study provides novel evidence that serum NfL levels during recovery from an mTBI are reflective of vulnerability to poorer outcomes in the event of a subsequent mTBI. We found that serum NfL levels at the time of the second impact were predictive of mTBI severity, with rats with high NfL displaying more observable neurological signs, and a greater rise in subsequent NfL levels when compared to rats with low NfL. Notably, the potentiated NfL increase phenomenon remained when the analysis was limited to rats with longer recovery periods (i.e. 7- and 14-days). While it is difficult to translate biological timeframes between rodents and humans, it is accepted that biological processes in rodents are relatively accelerated. As such, these findings indicate that this vulnerability may be present for several weeks or months after clinical concussion. Considered alongside emerging clinical evidence that serum NfL rises are a common but heterogeneous feature of mTBI and sport-related concussion, these novel findings suggest that monitoring NfL levels during recovery may be informative for individual return to duty, work or sport decisions.

      Commentary

      Background: Axonal injury is a prominent and potentially ongoing pathophysiological feature of mTBI. Serum NfL levels have been shown to be substantially but variably increased after mTBI. The potential for repeated mTBIs in short succession having cumulative effects is widely recognized but the mechanism is poorly understood.
      Translational significance: This study shows for the first time that serum NfL levels measured after mTBI might indicate the extent of cerebral vulnerability to a second mTBI. This finding provides strong initial evidence that NfL might be useful to help inform the safe to return to sport or duty after mTBI.

      Funding Statement

      This work was supported by funding from the Australian National Health and Medical Research Council (2020/GNT2002689).

      Acknowledgements

      The authors acknowledge the facilities and scientific and technical assistance of the National Imaging Facility (NIF), a National Collaborative Research Infrastructure Strategy (NCRIS) capability at Alfred research Alliance - Monash Biomedical Imaging (ARA-MBI), a Technology Research Platform at Monash University. All authors have read the journal's policy on disclosure of potential conflicts of interest and declare no conflicts. All authors have read the journal's authorship agreement and reviewed and approved the manuscript.

      References

      1. Faul M, Wald MM, Xu L, Coronado VG. Traumatic brain injury in the United States; emergency department visits, hospitalizations, and deaths, 2002-2006. 2010.

        • Meehan 3rd, WP
        • Mannix R.
        Pediatric concussions in United States emergency departments in the years 2002 to 2006.
        J Pediatr. 2010; 157: 889-893
        • Levin HS
        • Diaz-Arrastia RR.
        Diagnosis, prognosis, and clinical management of mild traumatic brain injury.
        Lancet Neurol. 2015; 14: 506-517
        • Hoge CW
        • McGurk D
        • Thomas JL
        • Cox AL
        • Engel CC
        • Castro CA.
        Mild traumatic brain injury in U.S. Soldiers returning from Iraq.
        N Engl J Med. 2008; 358: 453-463
        • Dretsch MN
        • Silverberg ND
        • Iverson GL.
        Multiple Past Concussions Are Associated with Ongoing Post-Concussive Symptoms but Not Cognitive Impairment in Active-Duty Army Soldiers.
        J Neurotrauma. 2015; 32: 1301-1306
        • Miller KJ
        • Ivins BJ
        • Schwab KA.
        Self-reported mild TBI and postconcussive symptoms in a peacetime active duty military population: effect of multiple TBI history versus single mild TBI.
        J Head Trauma Rehabil. 2013; 28: 31-38
        • Iverson GL
        • Gardner AJ
        • Terry DP
        • Ponsford JL
        • Sills AK
        • Broshek DK
        • et al.
        Predictors of clinical recovery from concussion: a systematic review.
        Br J Sports Med. 2017; 51: 941-948
        • Silverberg ND
        • Lange RT
        • Millis SR
        • Rose A
        • Hopp G
        • Leach S
        • et al.
        Post-concussion symptom reporting after multiple mild traumatic brain injuries.
        J Neurotrauma. 2013; 30: 1398-1404
        • Eisenberg MA
        • Andrea J
        • Meehan W
        • Mannix R
        Time Interval Between Concussions and Symptom Duration.
        Pediatrics. 2013; 132: 8-17
        • Kamins J
        • Bigler E
        • Covassin T
        • Henry L
        • Kemp S
        • Leddy JJ
        • et al.
        What is the physiological time to recovery after concussion? A systematic review.
        Brit J Sport Med. 2017; : 51
        • Vagnozzi R
        • Tavazzi B
        • Signoretti S
        • Amorini AM
        • Belli A
        • Cimatti M
        • et al.
        Temporal window of metabolic brain vulnerability to concussions: mitochondrial-related impairment–part I.
        Neurosurgery. 2007; 61 (discussion 88-9): 379-388
        • Giza CC
        • Prins ML
        • Hovda DA.
        It's Not All Fun and Games: Sports, Concussions, and Neuroscience.
        Neuron. 2017; 94: 1051-1055
        • Vagnozzi R
        • Signoretti S
        • Tavazzi B
        • Floris R
        • Ludovici A
        • Marziali S
        • et al.
        Temporal window of metabolic brain vulnerability to concussion: a pilot 1H-magnetic resonance spectroscopic study in concussed athletes–part III.
        Neurosurg. 2008; 62: 1286-1296
        • McCrory P
        • Meeuwisse W
        • Dvorak J
        • Aubry M
        • Bailes J
        • Broglio S
        • et al.
        Consensus statement on concussion in sport-the 5(th) international conference on concussion in sport held in Berlin, October 2016.
        Br J Sports Med. 2017; 51: 838-847
        • VE Johnson
        • Stewart W
        • Smith DH.
        Axonal pathology in traumatic brain injury.
        Exp Neurol. 2013; 246: 35-43
        • Hill CS
        • Coleman MP
        • Menon DK.
        Traumatic Axonal Injury: Mechanisms and Translational Opportunities.
        Trends Neurosci. 2016; 39: 311-324
        • Miller DR
        • Hayes JP
        • Lafleche G
        • Salat DH
        • Verfaellie M.
        White matter abnormalities are associated with chronic postconcussion symptoms in blast-related mild traumatic brain injury.
        Hum Brain Mapp. 2016; 37: 220-229
        • McDonald SJ
        • O'Brien WT
        • Symons GF
        • Chen Z
        • Bain J
        • Major BP
        • et al.
        Prolonged elevation of serum neurofilament light after concussion in male Australian football players.
        Biomark Res. 2021; 9: 4
        • Shahim P
        • Tegner Y
        • Marklund N
        • Blennow K
        • Zetterberg H.
        Neurofilament light and tau as blood biomarkers for sports-related concussion.
        Neurology. 2018; 90: e1780-e17e8
        • Zetterberg H
        • Blennow K.
        Fluid biomarkers for mild traumatic brain injury and related conditions.
        Nat Rev Neurol. 2016; 12: 563-574
        • Shahim P
        • Zetterberg H
        • Tegner Y
        • Blennow K.
        Serum neurofilament light as a biomarker for mild traumatic brain injury in contact sports.
        Neurology. 2017; 88: 1788-1794
        • Clarke GJB
        • Skandsen T
        • Zetterberg H
        • Einarsen CE
        • Feyling C
        • Follestad T
        • et al.
        One-Year Prospective Study of Plasma Biomarkers From CNS in Patients With Mild Traumatic Brain Injury.
        Front Neurol. 2021; 12643743
        • McDonald SJ
        • Piantella S
        • O'Brien WT
        • Hale MW
        • O'Halloran P
        • Kinsella G
        • et al.
        Clinical and blood biomarker trajectories after concussion: New insights from a longitudinal pilot study of professional flat-track jockeys.
        J Neurotrauma. 2022;
        • O'Brien WT
        • Pham L
        • Brady RD
        • Bain J
        • Yamakawa GR
        • Sun M
        • et al.
        Temporal profile and utility of serum neurofilament light in a rat model of mild traumatic brain injury.
        Exp Neurol. 2021; 341113698
        • Pham L
        • Shultz SR
        • Kim HA
        • Brady RD
        • Wortman RC
        • Genders SG
        • et al.
        Mild Closed-Head Injury in Conscious Rats Causes Transient Neurobehavioral and Glial Disturbances: A Novel Experimental Model of Concussion.
        J Neurotrauma. 2019; 36: 2260-2271
        • Pham L
        • Wright DK
        • O'Brien WT
        • Bain J
        • Huang C
        • Sun M
        • et al.
        Behavioral, axonal, and proteomic alterations following repeated mild traumatic brain injury: Novel insights using a clinically relevant rat model.
        Neurobiol Dis. 2020; 105151
        • Davis GA
        • Makdissi M
        • Bloomfield P
        • Clifton P
        • Echemendia RJ
        • Falvey EC
        • et al.
        International consensus definitions of video signs of concussion in professional sports.
        Br J Sports Med. 2019; 53: 1264-1267
        • Ah Kim H
        • Semple BD
        • Dill LK
        • Pham L
        • Dworkin S
        • Zhang SR
        • et al.
        Systemic treatment with human amnion epithelial cells after experimental traumatic brain injury.
        Brain Behav Immun Health. 2020; 5100072
        • Andersson JL
        • Skare S
        • Ashburner J.
        How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging.
        Neuroimage. 2003; 20: 870-888
        • Tournier JD
        • Smith R
        • Raffelt D
        • Tabbara R
        • Dhollander T
        • Pietsch M
        • et al.
        MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation.
        Neuroimage. 2019; 202116137
        • Zamani A
        • O'Brien TJ
        • Kershaw J
        • Johnston LA
        • Semple BD
        • Wright DK.
        White matter changes following experimental pediatric traumatic brain injury: an advanced diffusion-weighted imaging investigation.
        Brain Imaging Behav. 2021; 15: 2766-2774
        • Rota M
        • Antolini L
        • Valsecchi MG.
        Optimal cut-point definition in biomarkers: the case of censored failure time outcome.
        BMC Med Res Methodol. 2015; 15: 24
        • Reyes J
        • Mitra B
        • Makdissi M
        • Clifton P
        • Nguyen JVK
        • Harcourt P
        • et al.
        Visible Signs of Concussion and Cognitive Screening in Community Sports.
        J Neurotrauma. 2022; 39: 122-130
        • Shultz SR
        • McDonald SJ
        • Corrigan F
        • Semple BD
        • Salberg S
        • Zamani A
        • et al.
        Clinical Relevance of Behavior Testing in Animal Models of Traumatic Brain Injury.
        J Neurotrauma. 2020; 37: 2381-2400
        • DeWitt DS
        • Hawkins BE
        • Dixon CE
        • Kochanek PM
        • Armstead W
        • Bass CR
        • et al.
        Pre-Clinical Testing of Therapies for Traumatic Brain Injury.
        J Neurotrauma. 2018; 35: 2737-2754
        • Weil ZM
        • Gaier KR
        • Karelina K.
        Injury timing alters metabolic, inflammatory and functional outcomes following repeated mild traumatic brain injury.
        Neurobiol Dis. 2014; 70: 108-116
        • Prins ML
        • Alexander D
        • Giza CC
        • Hovda DA.
        Repeated mild traumatic brain injury: mechanisms of cerebral vulnerability.
        J Neurotrauma. 2013; 30: 30-38
        • Longhi L
        • Saatman KE
        • Fujimoto S
        • Raghupathi R
        • Meaney DF
        • Davis J
        • et al.
        Temporal window of vulnerability to repetitive experimental concussive brain injury.
        Neurosurgery. 2005; 56 (discussion -74): 364-374
        • Grant DA
        • Serpa R
        • Moattari CR
        • Brown A
        • Greco T
        • Prins ML
        • et al.
        Repeat Mild Traumatic Brain Injury in Adolescent Rats Increases Subsequent beta-Amyloid Pathogenesis.
        J Neurotrauma. 2018; 35: 94-104