MRI Evidence for Altered Venous Drainage and
Intracranial Compliance in Mild Traumatic Brain Injury
Andreas Pomschar1, Inga Koerte1, Sang Lee2, Ruediger P. Laubender3, Andreas Straube4,
Florian Heinen5, Birgit Ertl-Wagner1, Noam Alperin2*
1 Institute of Clinical Radiology, University of Munich – Grosshadern Campus, Ludwig-Maximilians-University Munich, Munich, Germany, 2 Department of Radiology, Miller
School of Medicine, University Miami, Miami, Florida, United States of America, 3 Institute of Medical Informatics, Biometry, and Epidemiology (IBE), Ludwig-Maximilians-
University Munich, Munich, Germany, 4 Department of Neurology, Ludwig-Maximilians-University Munich, Munich, Germany, 5 Department of Pediatric Neurology and
Developmental Medicine, Dr. von Hauner’s Children’s Hospital, Ludwig-Maximilians-University Munich, Munich, Germany
Abstract
Purpose: To compare venous drainage patterns and associated intracranial hydrodynamics between subjects who
experienced mild traumatic brain injury (mTBI) and age- and gender-matched controls.
Methods: Thirty adult subjects (15 with mTBI and 15 age- and gender-matched controls) were investigated using a 3T MR
scanner. Time since trauma was 0.5 to 29 years (mean 11.4 years). A 2D-time-of-flight MR-venography of the upper neck was
performed to visualize the cervical venous vasculature. Cerebral venous drainage through primary and secondary channels,
and intracranial compliance index and pressure were derived using cine-phase contrast imaging of the cerebral arterial
inflow, venous outflow, and the craniospinal CSF flow. The intracranial compliance index is the defined as the ratio of
maximal intracranial volume and pressure changes during the cardiac cycle. MR estimated ICP was then obtained through
the inverse relationship between compliance and ICP.
Results: Compared to the controls, subjects with mTBI demonstrated a significantly smaller percentage of venous outflow
through internal jugular veins (60.9621% vs. controls: 76.8610%; p = 0.01) compensated by an increased drainage through
secondary veins (12.3610.9% vs. 5.563.3%; p,0.03). Mean intracranial compliance index was significantly lower in the mTBI
cohort (5.861.4 vs. controls 8.461.9; p,0.0007). Consequently, MR estimate of intracranial pressure was significantly higher
in the mTBI cohort (12.562.9 mmHg vs. 8.862.0 mmHg; p,0.0007).
Conclusions: mTBI is associated with increased venous drainage through secondary pathways. This reflects higher outflow
impedance, which may explain the finding of reduced intracranial compliance. These results suggest that hemodynamic
and hydrodynamic changes following mTBI persist even in the absence of clinical symptoms and abnormal findings in
conventional MR imaging.
Citation: Pomschar A, Koerte I, Lee S, Laubender RP, Straube A, et al. (2013) MRI Evidence for Altered Venous Drainage and Intracranial Compliance in Mild
Traumatic Brain Injury. PLoS ONE 8(2): e55447. doi:10.1371/journal.pone.0055447
Editor: Wang Zhan, University of Maryland, United States of America
Received May 8, 2012; Accepted January 2, 2013; Published February 6, 2013
Copyright: ! 2013 Pomschar et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported in part by National Institutes of Health award R01 NS05212. The funders had no role in study design, data collection and
analysis, decision to publish, or preparation of the manuscript.
Competing Interests: Noam Alperin is a shareholder in Alperin Noninvasive Diagnostics, Inc. A dedicated software tool 200 (MRICP version 1.4.35 Alperin
Noninvasive Diagnostics, Miami, FL) was used in this study. The algorithm used by the mentioned software tool is patented, and the patent owned by Alperin.
There are no further patents, products in development or marketed products to declare. This does not alter the authors’ adherence to all the PLOS ONE policies
on sharing data and materials.
* E-mail: nalperin@med.miami.edu
Introduction concentration and fatigue resulting in functional and financial
problems that impact the healthcare system and society in general.
Traumatic brain injury (TBI) affects over 1.4 million individuals The pathophysiology associated with symptoms reported after
annually in the United States alone [1]. The majority of TBI are mTBI still remains to be further elucidated. There is histological
classified as mild traumatic brain injury (mTBI) defined as a blunt evidence for diffuse morphological changes after mTBI in humans
head trauma resulting in transient confusion, disorientation, including microscopic axonal injury leading to Wallerian degen-
impaired or loss of consciousness lasting 30 minutes or less in eration, transaction, and microglial clusters [4,5]. Advanced MR
combination with a number of unspecific neurological and based imaging and spectroscopy techniques, such as functional
cognitive symptoms [1–3]. Despite the lack of cerebral abnormal- MRI (fMRI) [6,7], diffusion tensor imaging (DTI) [8–10], and MR
ities on conventional anatomical imaging such as CT or MRI, spectroscopy [11] have shown to be more sensitive in detecting
patients experience a great variety of long-term neurological alterations in the brain following mTBI. For example, increased
symptoms such as orthostatic hypotension, headache, disturbed fractional anisotropy (FA) and decreased radial diffusivity in the
white matter e.g. the corpus callosum were detected using DTI,
PLOS ONE | www.plosone.org 1 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
suggestive of axonal cytotoxic edema [8–10]. Recent fMRI studies Subjects and Methods
found a more disperse brain activation pattern with a pronounced
activation in the dorsolateral prefrontal cortex and the hippocam- Ethics Statement
pus after mTBI [6,7]. MRS demonstrated widespread cellular Approval from the Ethics Committee of the Medical Faculty at
metabolic dysfunction including a decrease of N-acetyl aspartate Ludwig-Maximilians-University, Munich, Germany was obtained
and an increase in total choline. These changes correlated with prior to the study and all study participants provided written
neuropsychological parameters after mTBI [11]. informed consent.
Investigations of the potential impact of mTBI on the
microvasculature in humans are scarce. Morphological changes Subjects
in the cerebral microvasculature were documented in a nonhuman Thirty subjects were included in the study. The subjects were
primate animal model of mTBI, where a persistent increase in recruited among colleagues and co-workers in our university
endothelial projections in the arterioles and venules throughout hospital and through an announcement in the hospital’s internal
the brain was found after lateral head acceleration [12]. Though, clinical bulletin. All subjects were questioned in detail regarding
these morphological changes cannot be directly visualized with general and neurological health and completed a detailed
current imaging methods, they may lead to alterations in the standardized questionnaire based on the criteria for mTBI of the
biomechanical properties of the cerebral microvasculature, which Centers for Disease Control (CDC). The questionnaire asked for
in sum, could affect cerebral hemodynamics [13]. history and description of head trauma, time since trauma,
A link between cerebral vascular compliance and venous symptoms following the trauma such as headache, nausea,
drainage is implied from differences in venous drainage patterns vomiting, dizziness, neck pain, retrograde or anterograde amnesia,
and venous outflow pulsatility between upright and supine fatigue, irritability to light and sound, duration of each symptom,
postures [14]. In the supine position, the internal jugular veins problems concentrating or learning difficulties after the trauma.
(IJVs) are the dominant pathway of cerebrovenous drainage, while Inclusion criteria were a mTBI even fulfilling the CDC criteria
in the upright posture venous drainage is shifted to secondary that occurred at least 6 months or longer prior to the study, full
pathways. These secondary venous pathways have been described recovery from the clinical signs and symptoms associated with
as early as 1957 by Batson [15]. Although secondary venous mTBI within one week of the trauma, absence of chronic and
drainage is highly variable [16,17] studies have consistently ongoing symptoms following the trauma, and an age above 18
identified three main non-jugular venous drainage pathways: (a) years. Inclusion criteria for the control cohort were an age and
internal vertebral venous plexus i.e. the epidural veins (EV), (b) gender match with a subject in the mTBI cohort and the absence
vertebral artery venous plexus i.e. the vertebral veins (VV), and (c) of any history of traumatic brain injury or whiplash trauma.
the more posterior deep cervical veins (DCV). Changes in venous Exclusion criteria for the mTBI cohort included a history of mTBI
drainage pathways are also associated with changes in intracranial less than 6 month prior to inclusion into the study. Exclusion
compliance and pressure [14,18]. Therefore, the aims of this study criteria for both cohorts consisted of a history of neurological or
were (1) to evaluate craniocervical venous drainage and (2) to psychiatric disorders, including learning disabilities or of other
assess potential changes in intracranial compliance following chronic medical conditions (including hypertension, cardiac
mTBI. Venous drainage patterns are assessed qualitatively with conditions, diabetes mellitus or cancer) or medication intake
MR venography and quantitatively with measurements of (other than oral contraceptives in female subjects) and MR-related
volumetric flow rates with MR velocity-encoded phase contrast contraindications, including cardiac pacemakers, other metallic
imaging. MRI estimates of intracranial compliance and pressure implants or claustrophobia. According to the guidelines of the
(MR-ICP) are obtained using an experimental non-invasive Centers for Disease Control and Prevention [1], mTBI is assumed
technique that estimates compliance and pressure from measure- if head injury resulting from blunt trauma, acceleration or
ments of arterial inflow, venous outflow and CSF flow to and from deceleration forces was reported and additionally one or more of
the cranial vault during the cardiac cycle [19]. the following conditions were associated with the head injury: any
period of observed or self-reported transient confusion, disorien-
Figure 1. 3D Maximum Intensity Projected MR venography images demonstrating cerebral venous drainage patterns in a control
subject (A) and in a subject with mTBI (B). The internal jugular veins are well visualized in the control subject as the dominant outflow channel,
while in the mTBI subject, the IJVs are not visualized and venous drainage occurs primarily through secondary veins, i.e., the epidural (filled arrows)
and vertebral veins (empty arrow), which are well visualized.
doi:10.1371/journal.pone.0055447.g001
PLOS ONE | www.plosone.org 2 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
Table 1. Absolute and relative arterial inflow and venous outflow volumetric flow rates through the primary and secondary
venous channels.
Age Gen. TCBF JVF JVF Secondary venous outflow (%) MRV
ID (Years) (mL/m) (mL/m) (%) DCV VV EV Total grading
MTBI_01 20 m 883 696 78.8 0.4 0.0 1.9 2.3 3
MTBI_02 22 f 966 564 58.4 8.0 5.5 1.4 14.9 5
MTBI_03 23 m 1020 718 70.4 0.0 1.3 6.0 7.3 3
MTBI_04 23 m 586 480 81.9 0.0 0.0 5.8 5.8 4
MTBI_05 23 f 797 484 60.7 12.4 1.7 0.0 14.0 4
MTBI_06 24 m 873 423 48.5 1.5 22.4 2.5 26.4 5
MTBI_07 25 m 694 355 51.2 3.7 2.3 2.8 8.9 4
MTBI_08 25 m 790 601 76.0 1.3 0.2 0.0 2.8 3
MTBI_09 27 m 1192 849 71.2 1.7 3.3 0.9 5.9 4
MTBI_10 27 m 704 381 54.2 4.8 1.0 11.3 17.1 5
MTBI_11 29 m 880 792 90.1 0.0 0.0 0.0 0.0 2
MTBI_12 29 m 791 355 44.9 12.0 0.0 2.4 14.4 5
MTBI_13 32 f 727 479 65.8 10.0 0.0 0.7 10.8 n.a.
MTBI_14 33 m 848 0 0.0 10.9 9.6 22.7 43.2 5
MTBI_15 49 f 826 511 61.9 10.1 0.0 0.2 10.3 4
CTR_01 18 m 911 599 65.7 3.7 2.3 3.9 9.9 4
CTR_02 22 f 575 382 66.4 5.0 2.9 1.0 8.9 4
CTR_03 23 m 688 490 71.2 3.7 0.0 1.9 5.6 n.a.
CTR_04 23 m 831 747 89.9 1.8 0.0 2.5 4.4 2
CTR_05 24 f 809 548 67.7 1.3 0.0 0.0 1.3 2
CTR_06 24 m 819 740 90.4 0.0 0.0 3.1 3.1 3
CTR_07 24 m 888 753 84.9 0.0 1.4 3.0 4.3 2
CTR_08 25 m 797 693 86.9 0.0 0.0 2.5 2.5 2
CTR_09 26 m 904 723 80.0 7.4 0.0 0.0 7.4 4
CTR_10 26 m 716 421 58.9 6.0 0.0 5.8 11.9 n.a.
CTR_11 27 m 842 706 83.8 2.8 0.5 0.0 3.3 2
CTR_13 30 m 591 444 75.2 2.8 0.0 0.0 2.8 3
CTR_12 32 f 687 453 66.0 1.2 0.0 0.0 1.2 2
CTR_14 32 m 917 774 84.3 7.5 0.8 0.5 8.8 4
CTR_15 48 f 718 583 81.2 3.3 0.0 3.2 6.5 3
Mean mTBI 838 513 60.9 5.1 3.2 3.9 12.3 4.0
CTR 779 604 76.8 3.1 0.5 1.8 5.5 2.8
p-value 0.44 0.22 0.01 0.03 0.004
JVF jugular venous flow, DCV = deep cervical veins, VV = vertebral veins, EV = epidural veins.
doi:10.1371/journal.pone.0055447.t001
tation, impaired consciousness, or loss of consciousness lasting 30 trauma and all subjects were asymptomatic at the time of our
minutes or less, any period of observed or self-reported dysfunction study. Time since trauma ranged from 6 months to 29 years (mean
of memory (amnesia) around the time of injury, observed signs of 11.4 years). The fifteen age- and gender-matched subjects with no
other neurological or neuropsychological dysfunction, such as history of trauma were scanned with the same protocol (4 female;
nausea, vomiting, headache, dizziness, irritability, fatigue or poor range: 18–48 years; mean 27.067.2 years).
concentration. Fifteen subjects (4 female; range: 20–49 years;
mean 27.467.1 years) reported a mild traumatic brain injury Imaging Data Acquisition
either due to a car accident possibly combined with whiplash (3 Subjects were imaged in supine position using a 3 Tesla MR
subjects), or due to a fall related impact to the head (12 subjects). scanner (Verio, Siemens Healthcare, Erlangen, Germany) with a
Ten subjects reported unconsciousness lasting less than 30 min. All 12 channel phased array head and neck coil. The MRI study
15 subjects reported symptoms directly following the mTBI (e.g. protocol included conventional anatomical brain sequences such
headache, nausea, vomiting, dizziness, neck pain, retrograde and as FLAIR- and 3D T1-weighted images to rule out any structural
anterograde amnesia) lasting no longer than 2 days. mTBI related pathology. An axial 2D TOF MR venography was added to image
symptoms of all studied subjects resolved within 2 days after the the veins in the infratentorial and upper cervical region for
PLOS ONE | www.plosone.org 3 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
Figure 2. Examples of high and low velocity encoded phase contrast images from a control subject (left) and a subject with mTBI
(right). A–B: Flow compensated magnitude images showing the bright signal from blood vessels. The augmented venous outflow through the
epidural, vertebral veins (arrows) and the deep cervical veins (arrow heads) is well visualized. C–D: High-velocity encoding images used for
measurements of arterial inflow and venous outflow through the jugular veins. E–F: Low-velocity encoding images used for measurements of the
flow through the secondary channels (epidural, vertebral, and deep cervical veins) and the CSF flow. The lumen boundaries (red – arteries, blue- veins
and yellow and red – CSF and cord) were identified using the PUBS automated segmentation method.
doi:10.1371/journal.pone.0055447.g002
PLOS ONE | www.plosone.org 4 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
Figure 3. Derived volumetric flow waveforms obtained from a control subject (left) and a subject with mTBI (right). The total arterial
inflow (TCBF) and total venous flow though the jugular veins are shown in the top (Fig. 3A and 3B), respectively. The measured venous flow through
the epidural, deep cervical, and vertebral veins are shown in the bottom (Fig. 3C and 3D).
doi:10.1371/journal.pone.0055447.g003
assessment of the venous drainage pathways. Imaging parameters Volumetric flow rates through blood and CSF lumens were
included FoV of 160 mm, slice thickness of 2 mm, matrix size of obtained using a semi-automated pulsatility based segmentation
2566244, TR of 23 ms, TE of 5.4 ms, and a flip angle of 45 deg. (PUBS) method for improved reliability [21]. The PUBS method
Two retrospectively-gated velocity encoding (VENC) cine phase utilizes velocity information throughout the entire time series to
contrast scans, one with a high VENC of 70–90 cm/sec) and one identify blood vessels or CSF lumen pixels, which in turn, results
with a low VENC of 7–9 cm/sec were added to measure blood with a 4 folds increase in measurement reproducibility and
and CSF flow rates to and from the cranium. Imaging planes for increased measurement accuracy compared with manual delinea-
the blood and CSF flow measurements were placed at the height tion [21]. Time-dependent volumetric flow rate waveforms are
of the dens axis perpendicular to the internal carotid and vertebral obtained by integrating the flow velocities inside identified luminal
arteries, and at the mid C2 level, respectively, as described by Tain cross-sectional area over all 32 phase contrast images representing
et al. [20]. Imaging parameters of the phase contrast scans one cardiac cycle. Flow waveforms were obtained for each of the
included a FoV of 140 mm, matrix size of 2566179, slice thickness four main cervical arteries (left and right internal carotid artery
of 4 to 6 mm, and flip angle of 20 deg. Minimum TE and TR (LICA, RICA) and left and right vertebral artery (LVA, RVA)), for
were used for maximal temporal resolution. One average and two the primary venous pathways, the left and right internal jugular
views per segment were used to keep acquisition time around 1.5 vein (LIJV, RIJV), and for the following secondary venous
minutes per scan (approximately 90 cardiac cycles). pathways: the vertebral veins (VV), epidural veins (EV), and deep
cervical veins (DCV).
Assessment of Venous Drainage Patterns Total arterial blood flow, which is also the total cerebral blood
Venous drainage was assessed qualitatively and quantitatively. flow (tCBF), is obtained by summation of the flow in the four
Qualitative visual assessment was obtained using 3D-maximum arteries. Total jugular venous flow (tJVF) is defined as the sum of
intensity projection (MIP) models of the MRVs. The MRV source LIJV and RIJV. Secondary venous flow (SVF) is defined as the
images and the 3D reconstructed models were inspected to sum of the flow through the three secondary channels
determine the degree of secondary venous outflow by a board (VV+EV+DCV). Since total venous outflow is equal to tCBF,
certified neuroradiologist who was blinded to the subjects’ status primary and secondary venous flow are also given as percentage of
using the following scale: 1 - no; 2 - minimal; 3 - mild secondary the tCBF to account for inter-subject variability. In addition,
venous outflow and 4 - pronounced secondary venous outflow in cervical CSF stroke volume, i.e., the volume of CSF that flows
one of the three pathways (VV, EV or DCV); 5– pronounced back and forth between the cranium and the spinal canal, was
secondary venous outflow in two of the three pathways and 6 - obtained by time integration of the CSF flow waveforms.
maximum secondary venous outflow in all three pathways.
PLOS ONE | www.plosone.org 5 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
Figure 4. Derived volumetric flow rates and intracranial volume change waveforms obtained from a control subject (left) and a
subject with mTBI (right). The arterial-minus-venous (A–V) and the CSF flow waveforms are shown in Fig. 4A and 4B. The Intracranial volume
change during a cardiac cycle is shown in Fig. 4C and 4D, respectively. The CSF and the A–V waveforms are shown together to demonstrate the fact
that the craniospinal CSF flow dynamics is driven by the net trans-cranial blood flow. The CSF waveform follows the A–V waveform more closely in
the mTBI subject demonstrating the lower intracranial compliance compared to the matched control subject.
doi:10.1371/journal.pone.0055447.g004
Estimation of the Intracranial Compliance and Pressure based on the reported inverse relationship between compliance
Details of the derivation of the intracranial compliance and and ICP [23]. Volumetric blood and CSF flow rate waveforms
pressure have been previously described [19,22]. Briefly, based on and derived parameters were obtained using a dedicated software
the physical definition of compliance as a ratio of volume and tool (MRICP version 1.4.35 Alperin Noninvasive Diagnostics,
pressure changes, intracranial compliance is estimated from the Miami, FL).
ratio of the maximal (systolic) intracranial volume (ICVC) and
pressure fluctuations during the cardiac cycle (PTP-PG). The Volumetric Assessment of the Lateral Ventricles
change in intracranial volume (ICVC) is obtained from the The 3D T1 weighted images (MPRAGE) were used for
momentary differences between volumes of blood and CSF assessment of the lateral ventricular volumes using 3D Slicer (v.
entering and leaving the cranium as shown in equations 1 and 2, 3.6.3, Surgical Planning Laboratory, BWH, Boston, MA) by
placing two different regions of interests (ROI) in the left and right
lateral ventricles. A used defined threshold was used to identify the
DICVC(i)~½f :A(i){fV (i){fCSF (i)" Dt ð1Þ ventricular boundaries. Where necessary, the ventricular ROIs
were manually edited by a trained radiologist (A. P). The ventricle
P P volume was then quantified by multiplying the number of voxels
DICVC(i)~ ½fA(i){fV (i){fCSF (i)":Dt~0 ð2Þ inside the ventricular region and the voxel size.
cardiac cardiac
cycle cycle
Statistical Analysis
where fAis arterial inflow, fV is venous outflow, and fCSF is the Linear mixed effects regression models with rank-transformed
craniospinal CSF flow. Equation 2 states that in steady state, the dependent variables were used to test for differences between the
intracranial volume is on average constant over an entire cardiac mTBI and the matched subjects’ clusters as rank transformation is
cycle. This condition is used to account for the unmeasured beneficial for a small sample size. The intercept was allowed to
fraction of the total venous outflow. vary by the matched subjects (random effects term). A two-step
The pressure change is derived from the amplitude of the CSF procedure proposed by Conover and Iman [24] was used in this
pressure gradient (PG) waveform obtained using the Navier-Stokes work for the ranks conversion first followed by a parametric
relationships between derivatives of velocities and the pressure analysis (e.g., a linear mixed effects model) on the ranked data
gradient. An MRI equivalent of ICP (MRICP) is then obtained instead of on the original scales of the data. A non-parametric test
PLOS ONE | www.plosone.org 6 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
Table 2. MRI derived hydrodynamic parameters. signs of bleeding, brain concussion or enlarged ventricles were
seen. Volumes of the lateral ventricles were similar in the two
groups (mTBI 18.666.3 ml vs. controls 18.2610.5 ml; p = 0.351).
Age CSF SV PTP-PG ICVC MRICP In contrast, the MRV revealed significant differences between theGen.
two cohorts. Compared to control subjects, the mTBI cohort
ID (years) (ml) (mmHg/cm) (ml) (mmHg) demonstrated reduced drainage through the IJVs associated with
MTBI_01 20 m 0.70 0.053 0.58 14.9 an increased venous outflow through secondary pathways which
MTBI_02 22 f 0.51 0.054 0.41 16.4 was demonstrated by a median grading of 4 for the mTBI group
and 3 for the controls (p = 0.004). An example of cerebral venous
MTBI_03 23 m 0.72 0.043 0.49 10.6
drainage patterns in a control subject and in a subject with mTBI
MTBI_04 23 m 0.47 0.042 0.36 11.7
is shown in Fig. 1. The IJVs are well visualized in the control
MTBI_05 23 f 0.70 0.034 0.54 9.1 subject as the dominant outflow channel, while in the mTBI
MTBI_06 24 m 0.65 0.048 0.53 10.6 subject, the IJVs can hardly be distinguished from the dense
MTBI_07 25 m 0.30 0.030 0.53 9.6 network of draining secondary veins, i.e., the epidural and
MTBI_08 25 m 0.98 0.036 0.69 8.7 vertebral veins.
MTBI_09 27 m 0.66 0.053 0.64 12.1 The quantitative evidence for reduced drainage through IJVs
and increased secondary drainage is summarized in Table 1.
MTBI_10 27 m 0.21 0.053 0.31 16.7
Examples of magnitude and high and low velocity encoding phase
MTBI_11 29 m 0.54 0.047 0.49 14.9 images from a control and an mTBI subject are shown in Fig. 2.
MTBI_12 29 m 0.56 0.050 0.44 15.5 The blood and CSF lumen boundaries identified by the semi-
MTBI_13 32 f 0.60 0.030 0.33 11.2 automated pulsatility based segmentation method are overlaid on
MTBI_14 33 m 0.59 0.045 0.55 9.2 the phase images. Arterial flow toward the brain is shown in white,
while venous outflow is shown in black. Derived tCBF and total
MTBI_15 49 f 0.54 0.034 0.26 15.7
jugular volumetric flow rate waveforms obtained from the
CTR_01 18 m 0.84 0.054 0.82 10.4
segmented phase contrast images of a control subject and of a
CTR_02 22 f 0.41 0.031 0.50 9.3 subject with mTBI are shown in Fig. 3a and Fig. 3b, respectively.
CTR_03 23 m 0.60 0.049 0.57 9.1 Venous flow waveforms through the three secondary veins are
CTR_04 23 m 0.50 0.031 0.77 5.8 shown in Fig. 3c, and Fig. 3d, respectively. Despite a high inter-
CTR_05 24 f 0.40 0.022 0.27 10.7 individual variability in the amount of venous flow through the
secondary veins, a significantly higher fraction of venous outflow
CTR_06 24 m 0.69 0.069 0.74 10.4
occurs through these secondary veins in mTBI, compared to
CTR_07 24 m 0.63 0.023 0.56 7.4
control subjects (12.3611% vs. 5.563%; p,0.033). Consistently,
CTR_08 25 m 0.41 0.054 0.52 13.8 the relative drainage through the jugular veins was significantly
CTR_09 26 m 0.56 0.056 0.97 7.5 lower in mTBI (mTBI: 60.9621% vs. controls: 76.8610%;
CTR_10 26 m 0.53 0.041 0.64 9.2 p = 0.01). The differences in venous drainage are not related to the
CTR_11 27 m 0.60 0.058 0.85 7.6 magnitude of tCBF as there was no significant difference in the
mean tCBF between the two groups (mTBI: 8386147 ml/min vs.
CTR_13 30 m 0.41 0.030 0.44 6.7
controls: 7796112 ml/min; p = 0.439).
CTR_12 32 f 0.48 0.031 0.44 8.0
CTR_14 32 m 0.80 0.042 0.83 6.7 Intracranial Compliance and Pressure
CTR_15 48 f 0.21 0.023 0.27 8.6 Examples of measured net trans-cranial blood flow (arterial
Mean mTBI 0.58 0.043 0.48 12.46 inflow minus venous outflow or A–V) and cranio-spinal CSF
CTR 0.54 0.041 0.61 8.77 waveforms from a control and an mTBI subjects are shown in
p-value 0.66 0.07 0.0007 Figures 4a and 4b, respectively. As can be seen, the CSF waveform
‘‘follows’’ the A–V waveforms more closely in the mTBI case,
SV = stroke volume, PG = pressure gradient, ICVC = intracranial volume change, which is typical for low compliance. Since the CSF flow is driven
MRICP = MR derived intracranial pressure. by the A–V flow a tighter relationship indicates a less compliant
doi:10.1371/journal.pone.0055447.t002
intracranial compartment [25]. The waveforms of the intracranial
volume change during the cardiac cycle of the control and the
(Mann-Whitney-U) was applied to test group differences of the mTBI subjects are shown in Figures 4c and 4d, respectively. The
visual scores for venous drainage patterns and for the ventricular Intermediate hydrodynamic parameters such as maximal or peak-
volumes. Spearman’s rank correlation was used to test whether to-peak pressure gradient (PTP-PG) and maximal intracranial
any of the hemodynamics and hydrodynamics parameters were volume change (ICVC) and MR estimate of intracranial pressure
correlated with time post injury. A p-value of ,0.05 (two-sided) (MRICP) values are summarized in Table 2. While no statistically
was considered statistically significant. Statistical analyses were significant difference between the two groups was found for the
performed with R (version 2.12.2) and SPSS (version 20.0), PTP-PG (mTBI 0.0460.01 mmHg/cm vs. controls
respectively. 0.0460.01 mmHg/cm; p = 0.66) and the ICVC (mTBI
0.4860.1 ml vs. controls 0.6160.2 ml; p = 0.07), a trend toward
Results a lower maximal volume change was seen in the mTBI group. In
Venous Drainage contrast, the intracranial compliance index was significantly lower
(mTBI 5.861.4 vs. controls 8.461.9; p,0.0007), and consequent-
Conventional MR sequences did not demonstrate any differ-
ly MRICP was significantly higher (mTBI 12.562.9 mmHg vs.
ences between the two cohorts and no visible abnormalities such as
controls 8.862.0 mmHg; p,0.0007) in mTBI. Statistical signif-
PLOS ONE | www.plosone.org 7 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
icance was not derived by outliers. The higher MRICP values in supports the reliability of the findings of reduced intracranial
the mTBI cohort were all within the normative range. compliance in mTBI. In fact, effects size for the compliance index
Finally, Spearman’s rank correlation did not reveal association and the MRICP variables of 1.5 and 1.7, respectively are among
of the hemodynamics and hydrodynamics parameters with time the largest reported for group difference between old mTBI
since injury. Lowest P value of 0.13 was obtained for the relative injured patients and controls.
venous drainage through the smaller secondary channels. Another important observation is the persistence of the
hemodynamic and hydrodynamic changes following the mTBI.
Discussion A study employing MR diffusion techniques following moderate
traumatic brain injury demonstrated persistent changes in water
This study employs MR imaging methods to explore whether diffusivity even at 6 months following trauma [30]. A more recent
mTBI is associated with altered hemodynamics and hydrodynam- study by MacDonald et al. reports abnormal findings at a much
ics. Qualitative and quantitative assessments by MR venography higher rate than is expected by chance in scans performed 6 to 12
and MR velocity encoded arterial inflow and venous outflow months following enrolments in military personal with clinical
measurements demonstrate significant differences in venous diagnosis of mild, uncomplicated traumatic brain injury [31]. It is
drainage pattern between the asymptomatic mTBI subjects and therefore plausible that diffuse parenchymal changes that do not
the control cohort. mTBI is associated with a significantly smaller fully resolve over time, contribute to the reduced intracranial
fraction of the cerebral blood drainage through the IJVs, i.e., the compliance through changes in brain volume.
primary drainage pathway in the supine posture. On average, only Although potential mechanisms behind the altered hemody-
60.8% of the cerebral blood flow leave the brain though the IJVs namics and hydrodynamics found in mTBI cannot be established
in the mTBI subjects, compared to 76.8% in the matched subjects based on the data presented in this study, several factors may
who did not report mTBI. The IJV drainage fraction in the potentially contribute based on previously published data on
control subjects measured in this study is in excellent agreement animal models and humans. Potential mechanisms underlying the
with a previous study of healthy subjects that reported approxi- observed altered venous drainage and reduced intracranial
mately 75% of the cerebral blood drain through the IJVs in supine compliance in mTBI include:
posture and 40% in the upright posture [14]. As expected, the a) Structural changes in the brain microvasculature
smaller fraction of venous drainage through the IJVs in mTBI was following mild traumatic brain injury. It is known from
associated with increased venous drainage through secondary animal studies that long lasting changes occur in the endothelium
venous pathways, which drain the neurocranium in parallel to the and in the perivascular astrocytes of baboons after lateral head
IJVs [15,18]. Increased drainage through secondary veins usually acceleration [12]. Moreover, a recently published study in 12
occurs in the upright posture, partly due to complete or partial professional boxers found an impaired dynamic cerebral autoreg-
collapse of the IJVs. The reason for increased venous drainage ulation and a reduction in the cerebrovascular reactivity to
through the secondary veins in supine in subjects after mTBI is changes in carbon dioxide [32]. These changes could potentially
unclear at this time. Yet, similar changes in supine have previously be linked to the reduced overall intracranial compliance through
been reported in patients with idiopathic intracranial hypertension reduced vasculature compliance and to the increase in secondary
[26] and chronic migraine [27]. venous drainage.
The physiology of the cerebral venous system and its b) Post-inflammatory changes following mild traumatic
relationship to the intracranial hydrodynamics is still not fully brain injury. Increases in inflammatory cytokines (e.g. IL 2, IL
understood. The concept of venous hemodynamics is complicated 6 and TNF alpha) have been reported following mTBI. A recent
by the fact that veins are collapsible blood vessels, which are study by Jin et al. documented changes in inflammatory cell
characterized by marked changes in their cross sectional marker expression and cellular infiltration in a controlled cortical
configuration in response to even a slight change in transmural impact model in mice [33]. Migration of inflammatory cells is
pressure [18]. Increased secondary venous drainage in the supine known to mostly occur in the venous vasculature [34–36]. Recent
position has also been reported in idiopathic intracranial experimental studies have observed increased leukocyte-endothe-
hypertension [26], thus potentially linking increased secondary lium interactions in venules after TBI [34]. The resulting local
venous drainage in supine, which represents high impedance for inflammatory response may subsequently induce the formation of
venous outflow, with reduced compliance and higher ICP. An microthrombi occluding cerebral venules. Inflammatory responses
association between higher ICP and changes in venous flow might influence the endothelial cells and the local production of
characteristics has previously been documented in mechanical vasoactive substances such as NO and endothelin, which in turn
models of cerebral drainage and in patients [28]. can influence cerebral hemodynamics. An inflammatory reaction
A significantly lower intracranial compliance and higher MR- with a potentially increased in vessel walls rigidity may lead to or
derived estimate of ICP (MRICP) were found in the mTBI cohort. contribute to the observed altered venous hemodynamics and a
The mean value of 8.8 mmHg found in the control group is in lowered intracranial compliance. Yet, it is important to note that
excellent agreement with a mean value of 9.6 mmHg previously altered venous drainage is not specific to mTBI. Recent reports
reported in a study of 23 healthy young adults [29]. On average, documented altered venous drainage, without a change in
the MR-estimate of ICP in mTBI was 12.5 mmHg, which is intracranial compliance, in migraine headaches and multiple
approximately 3.7 mmHg higher than the mean MRICP value sclerosis [27,37].
found in the control group. Interestingly, neither the maximal The main limitation of the study is related to the self-reporting
volume change nor the pressure gradient changes were statistically used for including the subjects with mTBI. Therefore, there is a
different between the two cohorts. Yet, the ratio of these lack of objective verification of the nature of the mTBI. However,
parameters, i.e., the compliance index, did reach a strong the inclusion criteria were based on the guidelines and conceptual
statistically significant difference (p = 0.0007) between the two definition developed by the Centers for Disease Control and
cohorts. Statistically, the variability of the ratio of two other Prevention [1]. Another limitation relates to the studys cross-
parameters is larger than the individual variability. The fact that sectional design and the large heterogeneity of time span since the
the intracranial compliance index, did reach statistical significance trauma. This study design does not provide information regarding
PLOS ONE | www.plosone.org 8 February 2013 | Volume 8 | Issue 2 | e55447
Altered Venous Drainage and Compliance in MTBI
the timing and the rate at which these changes occur following the precise mechanism of these alterations remains to be elucidated.
trauma. However, the persistence of the changes found in this However, these results suggest that MR measures of compliance or
study is consistent with existing literature demonstrating that mild ICP are potentially a sensitive marker for detecting hemodynamic
TBI leads to irreversible changes. These preliminary results and hydrodynamic changes associated with mTBI.
warranted further investigations including larger cohorts and
specifically designed to address the time evolution and magnitude Acknowledgments
of these alterations in relations to the elapsed time after trauma
and trauma severity, respectively. This work is part of Andreas Pomschar doctoral thesis
In conclusion, the current study provides imaging-based
evidence of a chronically reduced cerebral venous drainage Author Contributions
through the internal jugular veins, an increased drainage in Statistical analyses: RPL. Participated in the setting up of the study: FH.
secondary venous channels with a concomitantly mild decreased Conceived and designed the experiments: NA BE-W AP. Performed the
intracranial compliance and mild increased intracranial pressure experiments: AP BE-W NA. Analyzed the data: AP SL. Contributed
in supine subjects with a history of mTBI. As all mTBI subjects reagents/materials/analysis tools: NA SL. Wrote the paper: NA AS IK B-
were free of symptoms at the time of the scan, the observed EW FH.
changes seem to have no apparent impact on brain function. The
References
1. National Center for Injury Prevention and Control (2003) Report to Congress 20. Tain RW, Ertl-Wagner B, Alperin N (2009) Influence of the compliance of the
on Mild Traumatic Brain Injury in the United States: Steps to Prevent a Serious neck arteries and veins on the measurement of intracranial volume change by
Public Health Problem. Atlanta, GA: Centers for Disease Control and phase-contrast MRI. J Magn Reson Imaging 30: 878–883.
Prevention; 2003. 21. Alperin N, Lee SH (2003) PUBS: pulsatility-based segmentation of lumens
2. Jennett B (1996) Epidemiology of head injury. J Neurol Neurosurg Psychiatry conducting non-steady flow. Magn Reson Med 49: 934–944.
60: 362–369. 22. Miyati T, Mase M, Kasai H, Hara M, Yamada K, et al. (2007) Noninvasive
3. Kibby MY, Long CJ (1996) Minor head injury: attempts at clarifying the MRI assessment of intracranial compliance in idiopathic normal pressure
confusion. Brain Inj 10: 159–186. hydrocephalus. J Magn Reson Imaging 26: 274–278.
4. Blumbergs PC, Scott G, Manavis J, Wainwright H, Simpson DA, et al. (1994) 23. Marmarou A, Shulman K, LaMorgese J (1975) Compartmental analysis of
Staining of amyloid precursor protein to study axonal damage in mild head compliance and outflow resistance of the cerebrospinal fluid system. J Neurosurg
injury. Lancet 344: 1055–1056. 43: 523–534.
5. Povlishock JT, Katz DI (2005) Update of neuropathology and neurological 24. Conover WJ, Iman RL (1981) Rank Transformations as a Bridge Between
recovery after traumatic brain injury. J Head Trauma Rehabil 20: 76–94. Parametric and Nonparametric Statistics. The American Statistician 35: 124.
6. Slobounov SM, Zhang K, Pennell D, Ray W, Johnson B, et al. (2010) Functional 25. Tain RW, Alperin N (2009) Noninvasive intracranial compliance from MRI-
abnormalities in normally appearing athletes following mild traumatic brain based measurements of transcranial blood and CSF flows: indirect versus direct
injury: a functional MRI study. Exp Brain Res 202: 341–354. approach. IEEE Trans Biomed Eng 56: 544–551.
7. Zhang K, Johnson B, Pennell D, Ray W, Sebastianelli W, et al. (2010) Are 26. Alperin N, Lee SH, Mazda M, Hushek SG, Roitberg B, et al. (2005) Evidence
functional deficits in concussed individuals consistent with white matter for the importance of extracranial venous flow in patients with idiopathic
structural alterations: combined FMRI & DTI study. Exp Brain Res. intracranial hypertension (IIH). Acta Neurochir Suppl 95: 129–132.
8. Chu Z, Wilde EA, Hunter JV, McCauley SR, Bigler ED, et al. (2010) Voxel- 27. Koerte IK, Schankin CJ, Immler S, Lee S, Laubender RP, et al. (2011) Altered
based analysis of diffusion tensor imaging in mild traumatic brain injury in Cerebrovenous Drainage in Patients With Migraine as Assessed by Phase-
adolescents. AJNR Am J Neuroradiol 31: 340–346. Contrast Magnetic Resonance Imaging. Invest Radiol.
9. Mayer AR, Ling J, Mannell MV, Gasparovic C, Phillips JP, et al. (2010) A
28. Piechnik SK, Czosnyka M, Richards HK, Whitfield PC, Pickard JD (2001)
prospective diffusion tensor imaging study in mild traumatic brain injury.
Cerebral venous blood outflow: a theoretical model based on laboratory
Neurology 74: 643–650.
simulation. Neurosurgery 49: 1214–1222; discussion 1222–1213.
10. Wilde EA, Ramos MA, Yallampalli R, Bigler ED, McCauley SR, et al. (2010)
29. Alperin N, Lee SH, Sivaramakrishnan A, Lichtor T (2005) Relationship between
Diffusion tensor imaging of the cingulum bundle in children after traumatic
total cerebral blood flow and ICP measured noninvasively with dynamic MRI
brain injury. Dev Neuropsychol 35: 333–351.
technique in healthy subjects. Acta Neurochir Suppl 95: 191–193.
11. Govind V, Gold S, Kaliannan K, Saigal G, Falcone S, et al. (2010) Whole-brain
proton MR spectroscopic imaging of mild-to-moderate traumatic brain injury 30. Kumar R, Husain M, Gupta RK, Hasan KM, Haris M, et al. (2009) Serial
and correlation with neuropsychological deficits. J Neurotrauma 27: 483–496. changes in the white matter diffusion tensor imaging metrics in moderate
12. Maxwell WL, Whitfield PC, Suzen B, Graham DI, Adams JH, et al. (1992) The traumatic brain injury and correlation with neuro-cognitive function.
cerebrovascular response to experimental lateral head acceleration. Acta J Neurotrauma 26: 481–495.
Neuropathol 84: 289–296. 31. Mac Donald CL, Johnson AM, Cooper D, Nelson EC, Werner NJ, et al. (2011)
13. Alperin N, Varadarajalu B, Fisher C, Lichtor T (2002) Long-lasting Changes in Detection of blast-related traumatic brain injury in U.S. military personnel.
Blood and CSF Flow Dynamics Following Mild Traumatic Brain Injury. Proc N Engl J Med 364: 2091–2100.
Intl Soc Mag Reson Med 2. 32. Bailey DM, Jones DW, Sinnott A, Brugniaux JV, New KJ, et al. (2012) Impaired
14. Alperin N, Lee SH, Sivaramakrishnan A, Hushek SG (2005) Quantifying the cerebral haemodynamic function associated with chronic traumatic brain injury
effect of posture on intracranial physiology in humans by MRI flow studies. in professional boxers. Clin Sci (Lond).
J Magn Reson Imaging 22: 591–596. 33. Jin X, Ishii H, Bai Z, Itokazu T, Yamashita T (2012) Temporal changes in cell
15. Batson OV (1957) The vertebral vein system. Caldwell lecture, 1956. marker expression and cellular infiltration in a controlled cortical impact model
Am J Roentgenol Radium Ther Nucl Med 78: 195–212. in adult male C57BL/6 mice. PLoS One 7: e41892.
16. Doepp F, Schreiber SJ, von Munster T, Rademacher J, Klingebiel R, et al. 34. Schwarzmaier SM, Kim SW, Trabold R, Plesnila N (2010) Temporal profile of
(2004) How does the blood leave the brain? A systematic ultrasound analysis of thrombogenesis in the cerebral microcirculation after traumatic brain injury in
cerebral venous drainage patterns. Neuroradiology 46: 565–570. mice. J Neurotrauma 27: 121–130.
17. Stoquart-Elsankari S, Lehmann P, Villette A, Czosnyka M, Meyer ME, et al. 35. Piccio L, Rossi B, Colantonio L, Grenningloh R, Gho A, et al. (2005) Efficient
(2009) A phase-contrast MRI study of physiologic cerebral venous flow. J Cereb recruitment of lymphocytes in inflamed brain venules requires expression of
Blood Flow Metab 29: 1208–1215. cutaneous lymphocyte antigen and fucosyltransferase-VII. J Immunol 174:
18. Schaller B (2004) Physiology of cerebral venous blood flow: from experimental 5805–5813.
data in animals to normal function in humans. Brain Res Brain Res Rev 46: 36. Kivioja J, Rinaldi L, Ozenci V, Kouwenhoven M, Kostulas N, et al. (2001)
243–260. Chemokines and their receptors in whiplash injury: elevated RANTES and
19. Alperin NJ, Lee SH, Loth F, Raksin PB, Lichtor T (2000) MR-Intracranial CCR-5. J Clin Immunol 21: 272–277.
pressure (ICP): a method to measure intracranial elastance and pressure 37. Ertl-Wagner B, Koerte I, Kumpfel T, Blaschek A, Laubender RP, et al. (2012)
noninvasively by means of MR imaging: baboon and human study. Radiology Non-specific alterations of craniocervical venous drainage in multiple sclerosis
217: 877–885. revealed by cardiac-gated phase-contrast MRI. Mult Scler 18: 1000–1007.
PLOS ONE | www.plosone.org 9 February 2013 | Volume 8 | Issue 2 | e55447