Professional Documents
Culture Documents
Hypertonic Saline Attenuates Cord Swelling and Edema in Experimental Cord Injury - Experimental (2009) PDF
Hypertonic Saline Attenuates Cord Swelling and Edema in Experimental Cord Injury - Experimental (2009) PDF
Hypertonic Saline Attenuates Cord Swelling and Edema in Experimental Cord Injury - Experimental (2009) PDF
Objective: To use magnetic resonance imaging (MRI) to char- extending rostrally and caudally was demonstrated by MRI. Also,
acterize secondary injury immediately after spinal cord injury at this time point, a hypointense core was identified on T1, PD,
(SCI), and to show the effect of hypertonic saline on MRI indices and T2 weighted images. Over time this hypointense core reduced
of swelling, edema, and hemorrhage within the cord. in size and in some animals was no longer visible by 8 hours after
Design: A prospective, randomized, placebo-controlled study. SCI, although histopathology demonstrated presence of red blood
Setting: Research laboratory. cells. A prominent ring of T2-weighted image hyperintensity,
Subjects: Twelve adult Long-Evans female rats. characteristic of edema, surrounded the hypointense core. At the
Interventions: Rats underwent a unilateral 12.5 mm SCI at lesion center, this rim of edema occupied the entire unilateral
vertebral level C5. Animals were administered 0.9% NaCl (n ⴝ 6) injured cord and in all animals extended to the contralateral side.
or 5% NaCl (n ⴝ 6) at 1.4 mL/kg intravenously every hour starting Administration of HS resulted in increased serum [Na], attenua-
30 minutes after SCI. Immediately after SCI, rats were placed in a tion of cord swelling, and decreased volume of hypointense core
4.7T Bruker MRI system and images were obtained continuously and edema at the last time points.
for 8 hours using a home-built transmitter/receiver 3 cm Helm- Conclusions: We were able to use MRI to detect rapid and
holtz coil. Rats were killed 8 hours after SCI. acute changes in the evolution of tissue pathophysiology, and
Measurements and Main Results: Quantification of cord swell- show potentially beneficial effects of hypertonic saline in acute
ing and volumes of hypointense and hyperintense signal within cervical SCI. (Crit Care Med 2009; 37:2160 –2166)
the lesion were determined from MRI. At 36 minutes after SCI, KEY WORDS: nervous system trauma; diagnostic imaging; hyper-
significant swelling of the spinal cord at the lesion center and tonic solutions; critical care; sodium
S pinal cord injury (SCI) is a de- and the estimated lifetime costs reach sult is a rapid improvement of arterial
bilitating and a costly condi- U.S. $1.7–3.1 million (1). We recently pressure and cardiac output. In cerebral
tion. In human SCI, injuries to studied a rodent model of cervical SCI edema, HS lowers intracranial pressure
the cervical region of the spi- using magnetic resonance imaging (MRI) by establishing an osmotic gradient be-
nal cord are the most common (52.4%), to monitor and quantify lesion develop- tween the intracellular and intravascular
ment over a 3-week time course (2, 3). space. In addition, improved cerebral
From this study, we hypothesized that blood flow and increased delivery of oxy-
*See also p. 2306. MRI would also be a valuable tool for gen cause a compensatory vasoconstric-
From the Brain and Spinal Injury Center (YSN, JCB, assessment of early secondary injury tion and a reduction in cerebral blood
MSB), Department of Neurological Surgery, University
of California, San Francisco, CA; Heart and Lung Re- events in the acute phase of SCI and that volume, which further lowers intracra-
search Institute (GM), Departments of Medical Virology, MRI would provide information that nial pressure (6). There have been few
Immunology and Medical Genetics (CAT) and Radiology could be used for therapeutic interven- investigations on the use of HS in SCI,
(PS), The Ohio State University, Columbus, OH. tion assessment as has been suggested in and although most studies report positive
Supported by funds from the National Institutes of
Health (NS-31193 and 38079), the New York State human SCI (4). The current investigation effects of HS on behavioral and his-
Center of Research Excellence (CO 19772), and The demonstrates the effects of hypertonic sa- topathologic outcomes, this has not led
Ohio State University, College of Medicine. line (HS) on lesion development depicted to widespread use of HS in human or
Work performed at Brain and Spinal Injury Center, by MRI during the acute phase of unilat- experimental SCI (13–16). HS seems a
Department of Neurological Surgery, University of Cal-
ifornia, San Francisco, California and Department of
eral cervical SCI. promising addition to a combinatorial
Neuroscience and Department of Radiology, The Ohio HS has been investigated for resusci- treatment strategy in SCI, and further
State University, Columbus, Ohio. tation in traumatic shock and for reduc- research is required to examine its poten-
The authors have not disclosed any potential con- ing intracranial pressure to treat cerebral tial beneficial role in SCI.
flicts of interest.
For information regarding this article, E-mail:
edema after traumatic brain injury (5– We recently showed that MRI is a valu-
michael.beattie@ucsf.edu 12). HS mobilizes free water from the able imaging modality to assess temporal
Copyright © 2009 by the Society of Critical Care intracellular into the extracellular space evolution of SCI and to distinguish dif-
Medicine and Lippincott Williams & Wilkins by osmotic force and reduction of periph- ferent severities of cervical SCI in rats
DOI: 10.1097/CCM.0b013e3181a05d41 eral vascular resistance. In shock, the re- (2). In that study, quantification of cord
RESULTS
All animals survived the surgical pro- Figure 1. Representative T1-weighted magnetic resonance images at the level of the lesion epicenter.
cedures and 8-hour continuous MRI pro- Consecutive images taken every 60 minutes show lesion development over time from 36 minutes after
tocol. The time between injury and gen- injury (left) to 8 hours after injury (right) in a control animal (top row) and an animal that received
eration of the first image by MRI was not hypertonic saline (bottom row).
different among groups and was 36 ⫾ 4
minutes for both groups (range: 21–54
minutes). Body temperature during sur-
gery and MRI was not different among
groups. Administration of HS according
to our protocol resulted in a serum [Na]
of 152 ⫾ 1 mmol/L which was signifi-
cantly higher than the serum [Na] in
control animals (138 ⫾ 4 mmol/L) at the
end point of our study (p ⫽ 0.003).
Immediately after injury, asymmetry
of the spinal cord due to ipsilateral swell- Figure 2. Representative T2-weighted magnetic resonance images at the level of the lesion epicenter.
Consecutive images taken every 60 minutes show lesion development over time from 36 minutes after
ing, extending rostrocaudally well beyond
injury (left) to 8 hours after injury (right) in a control animal (top row) and an animal that received
the level of SCI, and an ipsilateral core of hypertonic saline (bottom row).
hypointense signal were detectable by
MRI in all animals of both groups (Fig. 1).
The rostrocaudal extent of the hypoin- in the control group. In these animals,
tense core was visible over a length of the volume of hyperintense signal was
three slices (3.3 mm). T2-weighted MRI significantly higher at 8 hours than dur-
showed a rim of hyperintense signal sur- ing the first 2 hours (Fig. 6; p ⫽ 0.001).
rounding the hypointense core (Fig. 2). In animals that received HS, this volume
Unlike the hypointense core that was lim- was significantly smaller when compared
ited to the ipsilateral side, the hyperin- with the control animals at hours 7 and 8
tense signal extended to the contralateral (p ⫽ 0.008 and p ⫽ 0.003, respectively).
side of the cord. The spread of hyperin- Figure 3. Examples of tracings of the regions of Histopathologic analysis showed se-
tense signal into the contralateral cord interest for determination of volume measure- vere disruption of the normal spinal cord
was particularly visible in the central area ments. An overlay of a T1- and proton density- cytoarchitecture and vasculature at 8
of the cord, around the central canal. The weighted image (left) from which the whole cord hours after acute unilateral SCI (Fig. 7).
rostrocaudal extent of the hyperintense and hypointense core are traced. A T2-weighted Hemorrhage was apparent in a radial pat-
signal was visible over a length of seven image (right) demonstrates thresholding and tern, with red blood cells accumulated
slices (7.7 mm). During the 8-hour tracing of the hyperintense area. along the tracts of penetrating arteries
course of our study, an interesting find- and veins (Fig. 7A). Closer to the lesion
ing, directly visible from MRI, was that in epicenter, large accumulations of red
four animals the hypointense core re- that received HS, cord volume was signif- blood cells were present throughout the
duced in severity (two animals in each icantly smaller than in control animals at ipsilateral cord, with most present in the
group) and in three of these animals (one all time points (p ⫽ 0.017). Quantifica- core of the lesion (Fig. 7B). Volume of
control, two HS) this hypointense core tion of the volume of hypointense signal hemorrhage determined from histopa-
had completely disappeared toward the throughout the cord over time (Fig. 5) thology was not different between the two
end point of our study. showed that in animals that received HS study groups (control: 4.1 ⫾ 0.4 mm3;
Figure 3 shows examples of the trac- this volume was significantly smaller at 8 HS: 4.3 ⫾ 0.3 mm3) and length over
ing method used for quantification of hours than during the first 6 hours (Fig. which hemorrhage was present was not
cord swelling and volumes of hypo- and 5A; p ⫽ 0.002). On examination of the different between the two groups (con-
hyperintense signal. Quantification of slices taken at the lesion epicenter, in trol: 4.4 ⫾ 0.2 mm; HS: 4.3 ⫾ 0.2 mm).
cord volumes over the 8-hour study pe- both groups the volume of hypointense Infiltration of the cord with red blood
riod (Fig. 4) demonstrated that in both signal was smaller at hours 7 and 8 com- cells was seen throughout the ipsilateral
groups, cord volume significantly in- pared with hours 1, 3–5, and 1– 6, respec- cord and in nine of 12 animals red blood
creased over time; cord volume at hour 1 tively (Fig. 5B; p ⬍ 0.001). cells were also present on the contralat-
was significantly lower than at hours 3– 8 Quantification of hyperintense signal eral side of the central canal (Fig. 7C). In
(p ⬍ 0.002). Furthermore, in animals showed an increase in volume over time these animals, red blood cells were
Figure 5. Volume of hypointense signal determined for seven consecutive slices (A) and for one slice at the level of lesion epicenter (B) over time. In A,
volume of hypointense signal throughout the cord was significantly smaller at 8 hours than at 1– 6 hours in animals that received hypertonic saline (p ⫽
0.002). In B, volume of hypointense signal at the level of lesion epicenter was smaller in both groups at 7 and 8 hours compared with hours 1, 3–5, and
1– 6, respectively (p ⬍ 0.001). Data are presented as mean ⫾ SEM.
Figure 7. A, Representative section from the lesion showing the radial pattern in which red blood cells infiltrate the spinal cord. B, Representative section
at the level of lesion epicenter. Severe disruption of normal anatomy and accumulation of red blood cells is clearly visible. C, Representative section showing
red blood cell accumulation in the ipsilateral right hemicord and contralateral accumulation of red blood cells in the central gray area dorsal and
ventrolateral to the central canal (arrows). D, Motor neuron counts in the right hemicord are significantly lower than in the left hemicord (p ⬍ 0.001).
Furthermore, within the right hemicord, motor neuron counts were significantly lower in the 2.76-mm length of cord around lesion epicenter than in the
rostral and distal ends of the right hemicord (p ⬍ 0.001). Within the left hemicord there was no significant difference in motor neuron count throughout
the lesion. Data are presented as mean ⫾ SEM.