Angelos G. Kolias, Clinical Lecturer in Neurosurgery, Addenbrooke’s Hospital & University of Cambridge, Cambridge, UK firstname.lastname@example.org
Peter J. Hutchinson, Professor of Neurosurgery, Addenbrooke’s Hospital & University of Cambridge, Cambridge, UK email@example.com
Franco Servadei, Professor of Neurosurgery, University Hospitals of Parma and Reggio Emilia & University of Parma, Italy firstname.lastname@example.org
Decompressive craniectomy (DC) refers to the practice of removing a large bone flap and opening the underlying dura in order to control brain swelling and raised intracranial pressure (ICP).1 The two most frequent indications for a DC are traumatic brain injury (TBI) and ischaemic stroke. This newsletter article will focus on the use of supratentorial DC following TBI.
Hemi-craniectomy or unilateral DC refers to the removal of a large fronto-temporo-parietal bone flap, whereas bifrontal DC refers to the removal of a bone flap extending from the floor of the anterior cranial fossa anteriorly to the coronal suture posteriorly and to the middle cranial fossa floor bilaterally. Wide opening of the dura is a necessary part of the procedure.2
Overall, a primary DC undertaken at the same time as evacuating a haematoma is the most frequent indication.4 A secondary DC, undertaken usually after a period of ICP monitoring due to difficulties with controlling the ICP and/or clinico-radiological deterioration, is a less frequent indication.
Several uncontrolled studies presenting encouraging results with the use of DC have been published, especially so since the 1990s. However, uncontrolled studies suffer from several types of bias, and this was the motivation for the launch of two investigator-initiated multi-centre randomised trials in early 2000s: the DECRA trial and the RESCUEicp trial.5,6 A third multi-centre randomised trial (RESCUE-ASDH) was launched in 2014.7
The DECRA study was conducted between 2002 and 2010 in Australia, New Zealand and Saudi Arabia and focused on the role of early secondary DC for patients with severe diffuse TBI and mild-to-moderate intracranial hypertension.8 In this study, 155 patients were enrolled within the first 72 h following TBI if their ICP exceeded 20 mmHg for >15 min (continuously or intermittently) within a 1-h period, and if they did not respond to optimised first-tier ICP-lowering interventions (e.g. sedation, normocapnia, osmotherapy etc). The two arms of the trial were bifrontal DC and standard medical management or standard medical management alone. Patients in the surgical arm had better ICP control, received fewer interventions for raised ICP, and had a reduced length of stay in the intensive care unit (ICU). However, the investigators observed a higher rate of unfavourable outcomes (death, vegetative state, severe disability) in surgical patients at 6 months (70% versus 51%; OR 2.21; 95% CI 1.14–4.26; p=0.02). The mortality was similar in both arms (19% in surgical vs. 18% in medical patients). Nevertheless, 27% of patients in the surgical arm had bilaterally unreactive pupils compared with only 12% in the medical arm. A post hoc adjustment for pupil reactivity at baseline, which was necessary as pupil reactivity is known to be a major prognostic indicator of outcome following TBI, revealed that the rate of unfavourable outcome was not significantly different between the two arms (adjusted OR 1.90; 95% CI 0.95–3.79). Overall, the DECRA study indicated that early neuro-protective bifrontal DC for mild-to-moderate intracranial hypertension is not superior to medical management for patients with diffuse TBI.
Contrary to the DECRA study, the RESCUEicp trial examined the effectiveness of DC (bifrontal or unilateral) as a last-tier treatment for patients with severe, sustained and refractory post-traumatic intracranial hypertension.9 Patients were randomised to DC or continuing medical therapy if their ICP exceeded 25 mmHg for at least 1 h and could not be controlled by first-tier and second-tier ICP-lowering interventions. Overall, 408 patients were enrolled from 52 centres in 20 countries. At 6 months, DC resulted in mortality that was 22% lower than that with medical management. However, DC was also associated with higher rates of vegetative state, lower severe disability (dependent on others for care), and upper severe disability (independent at home) than medical management. The rates of moderate disability and good recovery with surgery were similar to those with medical management. In a prespecified sensitivity analysis, favourable outcomes (prespecified as upper severe disability, moderate disability and good recovery) occurred in 42.8% of surgical patients and 34.6% of medical patients (p=0.12). At 12 months, the mortality difference did not change but surgical patients improved further; as a result favourable outcomes (upper severe disability or better) were observed in 45.4% of the surgical patients and 32.4% of medical patients (p=0.01).
The RESCUEicp trial had a number of important differences from DECRA: surgical technique (bifrontal or unilateral DC versus bifrontal DC only); ICP threshold (25 mmHg versus 20 mmHg); duration of refractory intracranial hypertension (at least 1 h versus 15 min); timing of randomisation (any time when inclusion criteria are met versus within 72 h after TBI only); inclusion of patients with evacuated or non-evacuated intracranial haematomas (allowed in RESCUEicp but not in DECRA). These differences explain the contrasting findings of the two trials. However, both trials are valuable and complement each other as one looked at secondary DC as a neuro-protective measure (DECRA), whereas the other examined secondary DC as a last-tier intervention (RESCUEicp).
Acute subdural haematomas are still considered the most lethal subtype of TBI and they are often associated with contusions and brain swelling. However, there is a paucity of high-quality evidence in the literature regarding the best surgical strategy (primary DC or craniotomy) for these patients and there is considerable variation in practice. For example, a recent international survey found that a higher proportion of neurosurgeons from other European countries (48/110; 44 %) as compared with UK/Irish neurosurgeons (29/138; 21 %) use primary DC in more than half of ASDH cases (p<0.001).10
The RESCUE-ASDH study was funded by the UK National Institute for Health Research (NIHR) and was launched in 2014 in order to compare primary DC (bone flap left out) with craniotomy (bone flap replaced) for severe TBI patients undergoing evacuation of an ASDH.7
The criteria which are being used to determine eligibility of individual patients are:
* Patients with additional lesions (e.g. intracerebral haemorrhage, contusions) can be included
Eligible patients are randomised intra-operatively after evacuating the ASDH. Patients with significant brain swelling preventing safe replacement of the bone flap are not suitable for randomisation and are being followed-up in the context of an observational cohort. The primary endpoint is functional outcome assessed with the extended Glasgow Outcome Scale (GOSE) at 12-months post-injury.
Almost 140 patients have been randomised in the trial so far. The trial is in the process of roll-out to several countries worldwide. Clinicians interested to collaborate are encouraged to visit http://www.rescueasdh.org/contact-us for further information.
However, the burden of TBI is much higher in LMICs and patients have to be treated for TBI despite the absence of evidence directly applicable to these countries. These are issues that, we feel, are worthy of attention and further study as part of efforts to improve global neurotrauma care.11
Several TBI subtypes associated with brain swelling and/or raised intracranial pressure can be managed with a DC. However, on the basis of evidence from multi-centre clinical trials, we can conclude that:
The RESCUE-ASDH trial is currently open and aims to define the role of primary DC for patients with acute subdural haematomas and swelling. The global neurosurgical community needs to consider the role of DC, cranioplasty and other decompressive procedures (e.g. floating or hinge craniotomy) not just in HICs but also, and perhaps more importantly, in LMICs due to a much greater TBI burden.