Guest post by Dr Peter Hutchinson, Professor of Neurosurgery; Dr Angelos Kolias, Clinical Lecturer in Neurosurgery; and Dr David Menon, Professor of Anaesthesia – all at the University of Cambridge and Addenbrooke’s Hospital, Cambridge, UK. It is a response to the recent LITFL post by Dr Alistair Nichol RESCUEicp and the Eye of the Beholder.
* beholderNOUN: literary, archaic. A person who sees or observes someone or something.
We welcome the ongoing discussion which is centred on the “acceptable” degree of disability and the choice of dichotomisation points for the extended Glasgow Outcome Scale (GOS-E) in trials in traumatic brain injury (TBI). We discussed some of these issues in the correspondence published in the NEJM1 and we provide the relevant excerpts below:
“The DECRA trial, as compared with the RESCUEicp trial, enrolled patients with a lower intracranial-pressure threshold (20 mm Hg vs. 25 mm Hg) for shorter intervals (15 minutes vs. 1 to 12 hours), after lower intensities of therapy (stage 1 interventions vs. stage 1 and 2 interventions), and within a shorter interval after injury (all patients enrolled within 72 hours after injury vs. 44% of patients enrolled >72 hours after injury). At enrolment, the populations also differed with respect to expected outcome; the requirement for stage 2 interventions increases the relative risk of death by 60%. Hence, at 6 months, the pooled mortality of 37.5% in the RESCUEicp trial versus 18.7% in the DECRA trial is unsurprising.
In addition, our primary analysis showed a significant between-group difference in the GOS-E distribution and a substantial reduction in mortality with surgery; this finding differed from that of the DECRA trial, in which mortality was similar in the two groups. The severity of injury in the RESCUEicp trial underpinned dichotomization in the prespecified sensitivity analysis at upper severe disability (independent at home) or better. Given the high expectation of a poor outcome, the use of a “conventional” dichotomy would be as inappropriate as the use of it in populations with mild TBI (in whom disability-free survival is often attainable). This approach is concordant with recent recommendations. Upper severe disability is a better outcome than a modified Rankin score of 4 (on a scale from 0 [no symptoms] to 6 [death]), the threshold that has driven the use of craniectomy in patients with ischemic stroke”.
This discussion explicitly acknowledges that the expected outcome is dependent on the severity of the initial injury. Consequently, our treatment choices, and the expectations that we have from them, need to be calibrated to realistic expectations of benefit.
We are also grateful that the blog post by Dr Nichol provides us with an opportunity to re-emphasise our strong belief that the perspective of patients and their families should be taken into account in such decisions. It is not for clinicians to unilaterally decide whether a given degree of disability is “acceptable” or otherwise – the person who needs to accept an outcome is the patient. Consequently, we believe that the indirect input of the patient (as best as is possible), and of families, is critical when determining the degree of acceptable disability, and consequently whether a craniectomy should be considered. We believe that the concept of shared decision making can play a fundamental role in this respect2.
In the context of decompressive craniectomy (DC) for TBI, DECRA and RESCUEicp together, provide information that aids such discussions with families. In such discussions, we would urge against using loaded terms such as “favourable” or “unfavourable” which are inevitably laden with our own value judgments. It would be far more desirable, we believe, to simply state that the best evidence we have suggests that:
- DC, when used before other treatment options have been exhausted, does not improve mortality or clinical outcome
- DC, as a rescue intervention when most other interventions have failed, reduces mortality by about 20% in refractory intracranial hypertension
- At 12 months, about 60% of these additional survivors would be at least independent at home. The rest would be dependent at home or not recover consciousness3.
We should then offer to provide details of these outcome categories to further clarify issues. Our experience is that, when presented with this information, some families chose to proceed to DC, and some do not.
It is critical, however, to also openly acknowledge what we do not know, since this is the only way for us to chart further research, refine management, and improve outcomes.
First, intracranial hypertension is not the sole driver of outcome. For example, the presence of large bilateral dorsolateral brainstem lesions or severe diffuse axonal injury are likely to be drivers of outcome which DC (and indeed, any ICP-lowering intervention) simply cannot modify. Unfortunately, early MRI studies that allow the exclusion of these pathologies with a sufficient degree of confidence are not currently feasible in most patients. Large international studies (see here) are seeking to determine patient characteristics and biomarkers that might allow more refined prognostication, and (hopefully) inform such decisions.
Second, we are unsure whether all of the disability that that follows craniectomy is exclusively related to the underlying severity of disease, or is also contributed to by the intervention. This may be generic (e.g. through deformation of brain tissue – see here), or only apply to some variant of it (e.g. non-division of the falx cerebri leading to pressure on the anterior crossing fibres of the corpus callosum). Ongoing follow-up studies that are planned in both DECRA and RESCUEicp will address some of these issues. These results may allow us to refine the intervention.
Finally, we need to accept and make it clear to families, that choosing the best course of action in these circumstances is imperfect. We also need to acknowledge that patients similar to those that were recruited to RESCUEicp have a high likelihood of substantial disability, regardless of where we set our bar. Moreover, there is evidence that many patients adapt to a level of substantial disability that they may have previously regarded as unacceptable4. Hence, the perspective of patients and their families should be driving treatment decisions, supplemented by the evidence from clinical trials and the experience of treating clinicians.
- Hutchinson PJ, Kolias AG, Menon DK. Craniectomy for Traumatic Intracranial Hypertension. New England Journal of Medicine. 2016; 375(24):2403-4. [pubmed]
- Muehlschlegel S, Shutter L, Col N, Goldberg R. Decision Aids and Shared Decision-Making in Neurocritical Care: An Unmet Need in Our NeuroICUs. Neurocritical Care. 2015; 23(1):127-30. [pubmed]
- Hutchinson PJ, Kolias AG, Timofeev IS, et al. Trial of Decompressive Craniectomy for Traumatic Intracranial Hypertension. The New England Journal of Medicine. 2016; 375(12):1119-30. [pubmed]
- Honeybul S, Janzen C, Kruger K, Ho KM. Decompressive craniectomy for severe traumatic brain injury: is life worth living?. Journal of Neurosurgery. 2013; 119(6):1566-75. [pubmed]