Page 54 - NNS 2014 Program & Abstract Book

A1-15

HIGH INCIDENCE OF DELAYED SEIZURES IN SEVERE TBI

PATIENTS

Tubi, M.A.

1

, Real, C.R.

1

, Dinov, I.

3

, McArthur, D.L.

1

, Toga, A.W.

2

,

Engel, J.

1

, Vespa, P.M.

1

University of California Los Angeles, Los Angeles, USA

2

University of Southern California, Los Angeles, USA

3

University of Michigan, Ann Arbor, USA

After Traumatic Brain Injury (TBI) early seizure susceptibility is well

recognized but human studies of incidence and progression of epi-

leptogenesis are limited. Animal models have demonstrated that MRI

biomarkers may predict seizure susceptibility after TBI. In this study

we sought to determine the incidence and potential predictors for the

development of delayed (beyond 7 days) seizures in humans.

Forty-six patients with severe TBI, enrolled from 2001–2011, were

contacted 2–10 years post-injury to evaluate incidence of seizures. 37

patients had acute volumetric T1-weighted MRIs. Principal compo-

nents analysis (HCPC) of acute injury factors to determine predict-

ability of delayed seizures was performed using the LONI pipeline,

featuring anatomic MRI evaluation of area, volume, shape index,

average, curvature, curvedness, and fractal dimension.

Mean age at injury was 36.9

–

17.3 years (9 females and 37 males).

Mean GCS was 5.5

–

3.3. CT lesions included 23 (50%) contusional

and 18 (39.1%) DAI plus contusion. 25/46 (52.2%) developed sei-

zures after TBI. Seizures developed 0–7 days post-injury in 12 pa-

tients (26.1%), 7–30 days post-injury in 6 patients (13.0%) and after

one month post-injury in 7 patients (15.2%). Of the 12 patients that

developed seizures acutely, 4 (8.7%) subsequently had seizures after

30 days post-injury. Multivariate analysis demonstrated GCS, acute

neurosurgical intervention, and acute seizures independently predicted

subacute seizures. Thalamic shape measurements, computed from

acute MRIs in an automated hierarchical clustering using HCPC,

predicted subacute seizures with limited accuracy of 0.67. A subse-

quent global analysis predicted subacute seizures with 0.76 accuracy.

Our results indicate a high incidence of seizures after severe TBI

both acutely and subacutely. Potential early predictors of subacute

seizures include GCS, surgical intervention, and acute seizures. Acute

morphometric features of injury did not accurately predict subacute

seizures. Subacute seizures occur for reasons that are not yet clear.

Key words

magnetic resonance imaging, seizure, traumatic brain injury

A1-16

DECOMPRESSIVE CRANIECTOMY IN TRAUMATIC BRAIN

INJURY: DETERMINING OPTIMAL FLAP SIZE FOR BET-

TER INTRACRANIAL PRESSURE CONTROL

Martel, P.

, Schur, S., Marcoux, J.

McGill University, Department of Neurosurgery, Montreal, CA

While controversial, the use of decompressive craniectomy (DC) for

the treatment of refractory ICP remains widely used in the neuro-

trauma setting. Indeed, while the

DECRA

trial reports DC may be

associated with more unfavorable outcomes, our institution, similarly

to the ongoing

RESCUEicp

trial, recommends large fronto-temporo-

parietal DC. Furthermore, the literature shows great discrepancies as

to what constitutes optimal craniectomy size, with varying methods

for reporting and measuring flap size, usual guideline being a minimal

diameter of 12 cm. Our study aims to identify optimal bone flap size

and clarify the way it is reported. All the cases of severe TBI requiring

DC at the Montreal General Hospital over the last ten years were

retrospectively reviewed and we identified thirty cases that required

such a procedure for pure cerebral swelling. We correlated the cra-

niectomy size (diameter, circumference, surface area) to different

clinical variables (hospital and ICU stays, GOS, ICP control, hyper-

tonic infusion). Thirty patients with severe TBI were identified (mean

age of 30

–

2 years, presenting GCS of 7.5

–

0.4), all requiring DC at

63

–

9 hours after trauma for refractory ICP. Eighty three percent had

an initial Marshall CT-score of 3. Better ICP control was achieved for

DCs with a ratio of flap circumference over skull hemicircumference

of more than 65% (over 96 hours post-operatively, p

<

0.05). A ten-

dency towards less post-operative hypertonic infusion was also found

(9.8

–

1.7 L vs 6.1

–

1.9L; p

=

0.11). In our study, only 40% of DC

diameter over 12 cm (usual guideline) made it to the larger circum-

ference ratio group, hence achieving better ICP control. According to

our results, this 12 cm diameter threshold might be insufficient. Fur-

thermore, it is impossible to measure it preoperatively since it is a

post-operative measurement. By standardizing craniectomy flap

measurement, we hope to provide an easy intraoperative guideline

(flap circumference) in order to ensure an adequate size craniectomy

and potentially better outcome of patients through better ICP control.

Key words

bone flap size, decompressive craniectomy, fronto-temporo-parietal

craniectomy, severe traumatic brain injury

A1-17

BRAIN TISSUE OXYGENATION AND 3, 6-MONTH NEU-

ROLOGICAL OUTCOME IN SEVERE TRAUMATIC BRAIN

INJURY

Puccio, A.M.

, Chang, Y.F., Shutter, L.A., Okonkwo, D.O.

University of Pittsburgh, Department of Neurological Surgery, Pitts-

burgh, PA

Brain tissue oxygenation (PbtO

2

) monitoring has been utilized in the

severe traumatic brain injury (sTBI) population as an

in vivo

tool to

detect oxygenation changes in the acute recovery phase. It has been

previously reported that the longer the time a patient experiences a

PbtO

2

of

£

15 torr, the greater the likelihood of death. The purpose of

this study is to assess PbtO

2

values and its relationship to 3 and 6-month

outcome in adult sTBI.

Under an approved IRB protocol, PbtO

2

data were prospectively

collected on sTBI patients (Glasgow Coma Scale (GCS) score

<

9)

during the acute 5 days following injury. Glasgow Outcome Scale (GOS)

score, a measurement of neurologic outcome, was assessed at 3 and

6-months from injury and dicotomized into poor (GOS 1-3) and favor-

able outcome (GOS 4-5). Statistical analyses were performed using a

logistic regression model controlling for age and initial severity of injury.

258 adult, sTBI patients, mean age(

–

SD) was 38 (

–

17) years, with

69% male and a median GCS of 6 with 3 and 6-month outcome were

included. ICU management included ICP management per Guidelines

of sTBI Management; however, PbtO

2

was not treated, just monitored.

Post-trauma day (PTD) 2 PbtO

2

data was chosen for analysis to avoid

insertional microtrauma and minimize variability of values seen

within the first 24 hours. Receiver operating characteristic curve

analysis resulted in cut-off values: minimum PbtO

2

<

16.03 and

maximum

<

37.15, significant for poor 3-month neurological outcome

[OR 2.46 (1.29, 4.68),

p

=

.004, OR 2.71 (1.37, 5.33),

p

=

.006 re-

spectively]. There was no significance for 6-month neurological out-

come at these same cutoffs; however, PTD2 PbtO

2

values in

combination for minimum

<

18.3, maximum

<

37.2 and average

<

24.3

had a OR 2.02 (1.11, 3.67),

p

=

.021.

A-22