Decompressive craniectomy for diffuse cerebral swelling after trauma: Long-term outcome and ethical considerations
AbstractBACKGROUND: There is currently much interest in the use of decompressive for the management of diffuse cerebral swelling after trauma. Although the use of the procedure may improve survival, some of those survivors may be left severely disabled. The aim of this study was to see whether severe disability can be predicted and discuss the difficult ethical issue that this raises. METHODS: This was a retrospective cohort subgroup analysis of those patients with severe head injury in Western Australia between 2004 and 2008 who had had a decompressive craniectomy for intractably raised intracranial pressure despite maximal medical management. RESULTS: Among a total of 1,786 adult neurotrauma patients admitted between 2004 and 2008, 74 patients required a bifrontal decompressive craniectomy for intractably raised intracranial pressure. After the application of Corticosteroid Randomization After Significant Head Injury (CRASH) trial collaborators? prediction model, predicted and observed outcomes were compared. The mean timing and median timing of surgery were 42 hours and 30 hours after hospital admission, respectively. The timing of decompressive craniectomy was inversely correlated to the severity of the head injury (Spearman's correlation coefficient = -0.251, p = 0.031). At 18-month follow-up, 16 patients were deceased, 3 were in a persistent vegetative state, and 10 were severely disabled. In contrast to these unfavorable outcomes, 35 patients had a good outcome and 10 were moderately disabled at 18 months. The discrimination of the CRASH prediction model was excellent (area under receiver-operating characteristic curve, 0.905; 95% confidence interval, 0.829-0.982; p = 0.001). CONCLUSION: Our data provide some evidence that the CRASH prediction model may help clinicians and families to make informed decision about the benefits and risks of decompressive craniectomy for diffuse cerebral swelling.