In this single centre retrospective cohort study, there was no difference in functional outcomes at 6 months post stroke between the group that had a DC within 48 hours of acute ischaemic stroke onset and the group that had a DC more than 48 hours after an acute ischaemic stroke. In a study of malignant middle cerebral artery infarction, Elsawaf et al. randomised 46 patients into two groups – one group had a DC after obvious neurological deterioration, and the second group had an “ultra-early” DC within 6 hours of presentation 6. Both functional outcome and mortality were superior in the ultra-early DC group. The HAMLET study enrolled 32 patients for decompressive hemicraniectomy up to 96 hours post stroke and compared their outcomes to best medical management 2. Compared to medical management alone, there was no benefit to a DC > 48 hours post stroke. Randomisation over 48 hours after stroke showed no benefit to surgical decompression on any outcome measure.
The functional outcome for the overall cohort was poor. At 6 months post stroke, the best functional outcome from the study population was a mRS score of 2 for 2/46 patients (4.4%), meaning that they could function independently but not carry out all the tasks they had done prior to their stroke. Fourteen out of 46 patients were dead (30.4%). Furthermore, 27/46 (58.7%) patients had a mRS score of 4–5, meaning that they were severely disabled and not independent with their activities of daily living. Five out of 46 patients (10.9%) had a mRS ≤ 3 at 6 months. This compares unfavourably to a pooled analysis from 7 trials of patients who had large hemispheric infarction. In that meta-analysis, 30% of patients who had a decompressive hemicraniectomy had a mRS ≤ 3 at 6 months 7.
In our multivariate analysis, patient age did not influence functional outcome. There were limited numbers of patients aged > 60 in our study. However, 8 patients in the early DC group and 5 patients in the late DC were aged > 60. The DESTINY II trial examined decompressive hemicraniectomies in patients over the age of 60. At 12 months post stroke, none of the patients in either the intervention or control group had a mRS score of 0–2 8. Given that studies of DC in traumatic brain injury were found to have similarly poor functional outcomes in older patients, it is common practice in our centre to proceed with DC in these patients only if they have good premorbid function 9.
Our study found no significant difference in 6-month outcomes between patients that had hemicraniectomy for anterior circulation stroke and those with a suboccipital craniectomy for posterior circulation stroke. Eleven patients had suboccipital DCs for malignant cerebellar oedema. At 6 months post stroke all of these patients were either dead or severely disabled. These outcomes compare unfavourably to the limited data available on outcomes from suboccipital decompressive craniectomy for acute ischaemic stroke. Lindeskog et al reported a favourable outcome (mRS 0–3) in 54% of 22 patients who had suboccipital decompressions 10. In a meta-analysis of 11 studies (with 283 patients) of suboccipital DC the pooled event rate for moderate-severe disability was 28% and 20% for mortality, indicating a favourable outcome in the majority of patients 5. In the group that had anterior circulation strokes there were more left-sided DCs performed - of the 36 hemicraniectomies, 23 were left-sided and 13 were right-sided. The reason for this is unclear. There appears to be insufficient evidence available to suggest that dominant hemisphere strokes have a worse outcome than non-dominant hemisphere strokes 11. From this review of the practice in our institution it does not appear that there is a selection bias against undertaking DCs in patients with dominant hemisphere strokes.
More patients in the early group had a thrombectomy prior to their decompressive craniectomy compared to the late group. This is likely the result of these patients already being in the hospital and having immediate access to neurosurgical services when required as opposed to patients requiring to be transferred into the hospital. In our analysis of DCs, a preceding thrombectomy did not influence 6-month functional outcome. This finding is not surprising given that despite improvements in mTICI scores in the patients who had thrombectomies in both early and late DC groups, this subgroup of patients obviously went on to develop malignant cerebral or cerebellar oedema requiring a craniectomy. In a single-centre experience from Germany the rate of DC went from 2.2% of all acute ischaemic strokes in 2009 to 1.9% in 2017 12. In that centre, the numbers of stroke thrombectomies considerably increased over that time frame and the authors suggested that mechanical thrombectomies have resulted in lower numbers of patients requiring DCs. Our data is somewhat in keeping with their experience. The number of DCs performed annually has varied but the numbers peaked in 2015 with 15 procedures. Between 2013 and 2018, there was a five-fold increase in the number of stroke thrombectomy procedures in our institution 13.
It should be noted that there has been a change in the recommendations for the management of cerebral oedema following ischaemic stroke during the study period and this may have influenced both the numbers of patients who underwent DC operations and the timing of these procedures. In 2014 the American Heart Association (AHA) guidelines stated that “ventriculostomy is recommended in obstructive hydrocephalus after a cerebellar infarct but should be followed or accompanied by decompressive craniectomy” 14. In 2018 this was changed to advise that ventriculostomy was still recommended in the case of obstructive hydrocephalus but a concomitant DC may not be necessary depending on factors such as infarct size and brainstem compression 1. For anterior circulation strokes the 2014 AHA guidelines said that the benefit of DC in patients > 60 was uncertain. The 2018 guidelines advised that in patients > 60 a DC within 48 hours should be considered given that it reduced mortality by close to 50%. After the study enrolment had ended in 2021, the European Stroke Organisation published guidelines on the management of space-occupying brain infarction. Based on moderate quality of evidence they recommend surgical decompression within 48 hours for patients under 60 with hemispheric infarction to reduce the risk of death or poor outcomes 15. They caution that a “careful discussion” is required with the patient or representatives before surgery to advise about the possibility of survival with significant disability. Regarding cerebellar decompression they advise that there is continued uncertainty as to whether this procedure reduces the risk of death or severe disability in patients with posterior fossa stroke.
This study has some limitations. It is retrospective and conducted over a 7-year period which has seen some changes in practice to the management of acute ischaemic stroke, e.g. the increasing number of endovascular thrombectomies. It is also a single-centre experience which may introduce a selection bias. The decision to undertake a DC procedure may have been influenced by individual practitioner bias – both from the physicians selecting which patients were appropriate to refer to our centre and from practitioners within our institution. The selection of these patients may also have been influenced by whether or not they were already in the hospital or how long it would take to transfer them to the institution.