GFAP (-CT: 0.1 ng/ml (IQR 0.0-1.7), n
=
31;
+
CT: 2.0 ng/ml (0.4-
3.3), n
=
30) and CRP (-CT: 0.5 ug/ml (IQR 0.3-1.7),
+
CT: 27.5 ug/ml
(IQR 6.3-50.0)) were significantly higher (p
<
0.0001, Mann-Whitney
U). Logistic regression of the combined GFAP-CRP model (GFAP
Odds Ratio (OR) 12.9, CRP OR 1.6, Nagelkerke R
2
(NR
2
) 0.87,
p
<
0.01) performed better than GFAP (OR 7.5, NR
2
0.54, p
<
0.01) or
CRP (OR 1.5, NR
2
0.70, p
<
0.01) alone in detecting
+
CT. Likewise
the receiver-operating characteristic (ROC) curve of the combined
model had a higher area under the curve (AUC) of 0.986 (p
<
0.0001)
compared to GFAP (AUC 0.891) or CRP (AUC 0.931) individually.
Our data suggests that GFAP and CRP are part of distinct physio-
logical cascades in response to injury (Pearson’s r 0.272, p
=
0.034),
and supports the development of a combined panel of biomarkers to
improve acute TBI detection.
Key words
clinical trial, human studies, traumatic brain injury
A4-13
PLASMA LIPIDOMIC TBI BIOMARKER PROFILES - TRANS-
LATION FROM MOUSE TO HUMAN
Crawford, F.C.
1,2
, Emmerich, T.
1,3
, Abdullah, L.
1,2
, Mouzon, B.C.
1,2
,
Evans, J.
1,2
, Reed, J.
1,2
, Crynen, G.
1,3
, Montague, H.
1,4
, Hart, A.
1
,
Gonzalez, A.
1
, Dretsch, M.
5
, Mullan, M.J.
1–3
1
Roskamp Institute, Sarasota, USA
2
James A. Haley Veterans Hospital, Tampa, USA
3
The Open University, Milton Keynes, UK
4
University of Cardiff, Cardiff, UK
5
National Intrepid Center of Excellence, Bethesda, USA
Traumatic brain injury (TBI), in particular mild TBI (mTBI) is a
major problem for military and civilian populations. An objective
panel of TBI biomarkers and related conditions would enable ap-
propriate medical management, may indicate ongoing pathogenic
processes, provide guidance in therapeutic development, and could
be used to monitor outcome and response to treatment. We have
developed a mouse model of single and repetitive mTBI that shows
progressive neuroflammation and neurobehavioral changes, charac-
terized through to 2 years post injury. One next step is to identify
potential peripheral biomarkers of injury. Phospholipids (PLs) such
as phosphatidylcholine (PC) and sphingomyelin (SM), play a pro-
minent role in neuronal processes including neurotransmitter release,
neurite outgrowth and synaptogenesis, and brain lipid metabolism is
disrupted in our preclinical TBI models. We have used our lipidomic
platform (in-source collision induced dissociation (sCID) with full
scan liquid chromatography/MS (LC/MS)) to generate a temporal
profile of plasma lipidomic changes in our mouse model. Plasma
profiling demonstrates significant TBI-dependent changes in lipid
profiles that persist years after the injury, including significant in-
creases in PC and SM. To determine the clinical relevance of these
findings we will correlate plasma lipidomic changes with brain li-
pidomic changes in this model, but we are also validating our
findings by investigating plasma lipid profiles in human TBI popu-
lations. Our pilot human data, from lipidomic profiling in a cohort of
soldiers pre and post-deployment, show plasma lipid changes in
those with a documented TBI that are consistent with those seen in
the mouse model.
Funding: CDMRP funding (W81XWH-10-1-0759), and VA Merit
funding; MOMRP funding and Roskamp Foundation.
Key words
lipidomic, military, plasma
A4-14
MULTI-ANALYTE BIOMARKER PANEL PREDICTS
FUNCTIONAL OUTCOME 6 MONTHS AFTER TRAUMATIC
BRAIN INJURY
Diaz-Arrastia, R.
1
, Yue, J.K.
2
, Wang, K.K.
3
, Cooper, S.R.
2
, Sorani,
M.D.
2
, Valadka, A.B.
4
, Okonkwo, D.O.
5
, Manley, G.T.
2
1
Center for Neuroscience and Regenerative Medicine, USUHS, Be-
thesda, USA
2
Brain and Spinal Injury Center, San Francisco General Hospital,
San Francisco, USA
3
Center for Neuroproteomics and Biomarkers Reserach, University of
Florida, Gainesville, USA
4
Seton Brain and Spine Institute, Austin, USA
5
University of Pittsburgh Medical Center, Pittsburgh, USA
Blood biomarkers show promise as tools to identify mild TBI (mTBI)
patients at risk of long-term functional deficits, but given the hetero-
geneity of mTBI, it is unlikely that a single biomarker will have suf-
ficient sensitivity and specificity to be useful. We ran a multi-analyte
panel of 87 biomarkers (HumanMap v.2.0, MyriadRBM, Austin,
Texas) on a cohort of 62 participants in the TRACK-TBI Pilot Study.
90% had mTBI (GCS 13-15) and 50% had normal cranial CT. Plasma
was collected
<
24 hours after injury and stored frozen until assayed.
We considered only biomarkers where
>
90% of samples provided
values higher than the Lower Limit of Quantitation. Biomarkers
with
>
30% of TBI samples outside (either higher or lower) than the
95% confidence interval for healthy controls were analyzed using
binary logistic regression, with full recovery 6 months after injury
(GOSE 8) as the dependent variable. 20 biomarkers from the Hu-
manMap v.2.0 panel were included in the anlysis. A model including
only age, admission GCS, and CT findings correctly classified only
72.1% of cases, and explained only a trivial fraction of variance (Cox
and Snell R
2
(CSR
2
)
=
0.020, Nagelkerke R
2
(NR
2
)
=
0.029). Adding
UCHL-1 and GFAP to the model did not significantly improve the
model (classification 72.9%, CSR
2
=
0.122, NR
2
=
0.174). Adding the
20 HumanMap v.2.0 biomarkers substantially improved the model
(classification 100%, CSR
2
=
0.690, NR
2
=
1.00). These findings are
preliminary and must be replicated in a larger, independent cohort, but
indicate that multiplex assays of biomarkers are potentially useful for
identifying patients with mTBI who fail to make a complete recovery.
Key words
GOSE, Luminex, multiplex immunoassay
A4-15
EXTRACELLULAR MATRIX BIOMARKERS ARE SEVER-
ITY DEPENDENT AND REGIONAL SPECIFIC IN EXPERI-
MENTAL DIFFUSE BRAIN INJURY
Griffiths, D.R.
1,2
, Addington, C.
3
, Adelson, P.D.
1–3
, Stabenfeldt, S.
3,4
,
Lifshitz, J.
1,2,5
1
Department of Child Health, University of Arizona College of
Medicine, Phoenix, USA
2
Barrow Neurological Institute at Phoenix Children’s Hospital,
Phoenix, USA
3
School of Biological and Health Systems Engineering, Arizona State
University, Tempe, USA
4
Neuroscience Program, Arizona State University, Phoenix, USA
5
Phoenix VA Healthcare System, Phoenix, USA
The extracellular matrix (ECM) provides structural support for neu-
ronal, glial and vascular components of the brain, particularly through
A-41
1...,63,64,65,66,67,68,69,70,71,72 74,75,76,77,78,79,80,81,82,83,...168