Decompressive Craniectomy in Diffuse Traumatic Brain Injury - DECRA trial

A must read paper published in New England Journal of Medicine, April 21, 2011.
The study was led by Prof DJ Cooper MD.
Funded by the National Health and Medical Research Council of Australia and others: DECRA Australian Clinical Trials.
DECRA - decompressive craniectomy for refractory intracranial hypertension in TBI.

Background: It is unclear whether decompressive craniectomy improves the functional outcome in patients with severe  TBI and refractory raised ICP.

Methods: In between December 2002 and April 2010, 155 adults with severe diffuse  TBI and intracranial hypertension that was refractory to first tier therapies were randomly assigned to undergo either bifrontotemporalparietal decompressive craniectomy or standard care.

The original primary outcome was unfavourable outcome (composite of death, vegetative state or severe disability) as evaluated on the GOS-E at 6 months after the injury.

The final primary outcome was the score on the GOS-E at 6 months.


Results:

Patients in the craniectomy group had less time with ICPs above the treatment threshold (p<0.001), fewer intervention for raised ICP and fewer days in ICU.

However patients undergoing craniectomy had worse scores on the GOS-E than those receiving standard care and greater risk of an unfavourable outcome.


Rates of death at 6 monts were similar in the craniectomy group (19%) and the standard care group (18%).


Conclusions:

In adults with severe diffuse TBI and refractory intracranial hypertension, early bifrontotemporoparietal decompressive craniectomy decreased intracranial pressure and length of stay in the ICU but was associated with more unfavourable outcomes.

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

For many years Neurosurgeons and intensivists have speculated that decompressive craniectomy in patients with severe TBI and refractory intracranial hypertension my improve clinical outcomes eventhough there have been no well conducted randomised controlled trials in its favour. DECRA was conducted by ANZICS CTG and studied  155 patients over 7 years. 
This randomized trial revealed unexpected results that  showed early decompressive craniectomy reduced ICP and the length of stay in the ICU but it was also associated with a greater risk for unfavourable outcome at 6 months compared with standard care. Rates of death didn't differ between groups, but scores on the GOS-E were lower in the surgical group  and there was a significant increase in risk, more than double for an unfavourable outcome on that same scale.

The findings differ from those of most nonrandomized studies and are contrary to the hypothesis. This surgical strategy is increasingly used at neurotrauma centres internationally.

The original primary outcome was unfavourable outcome on the GOS-E, a composite of death, vegetative state or severe disability. However, after the interim analysis in 2007, the primary outcome was revised to be the functional outcome at 6 months after injury on the basis of proportional odds analysis  of the GOS-E.

The editorial points out that most neurosurgeons wouldn't consider this aggressive strategy in patients who have increased ICP  for such a short time, in this study ICP more than 20mmHg for 15 minutes. In addition,  in screening of 3478 patients, only 155 patients enrolled in the trial suggests a selected population excluding both patients with mass lesions  and those whose intracranial pressure was successfully brought under control. 

For reasons that are not clear decompressive craniectomy appeared to convert survivors from a favorable outcome to an unfavorable outcome. Among many possible explanations are variations in surgical technique and unintended changes in brain physiology - swollen brain expansion outside the skull causing axonal stretch and injury of changes in cerebral blood flow. Other issues that may explain the results include heavy enrolment from a single centre or baseline imbalances  between groups. Even considering these effects, craniectomy was not shown to be beneficial.

The main lesson from this study is that surgical reduction of ICP by the technique that was used by the investigators does not necessarily result in better outcomes for patients and indeed appears to worsen them in at least some circumstances. The procedures should not be abandoned on the basis of these results. Rather, the risks and benefits of the decompressive craniectomy must be weight carefully and must work to define appropriate clinical settings for this procedure. Caution should be applied in the routine use of this strategy.
xoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxoxox

Further comments:
1. 3478 patients were screened over 7 years to enrol 155 patients. Shouldn't this trial have been stopped for futility?
2. Median ICP at randomization was 20 mmHg, did these patients even have intracranial hypertension?
3. ICP was managed for 15 minutes before surgery, is this clinical equipoise?
4. In this trial, decompressive craniectomy was performed at an average ICP of 22 mmHg for 30 minutes
5. There were 2.5x as many patients with bilateral fixed pupils in the surgical group. This between group difference was statistically significant (p=0.04).
6. After post-hoc adjustment for pupil reactivity at baseline, the between group differences were no longer significant for the score on the GOSE and for the risk of an unfavourable outcome.
7. the conclusion that decompressive craniectomy is associated with more unfavourable outcomes is highly misleading and not supported by the data.