Surgery for acute subdural haematoma: the value of pre-emptive decompressive craniectomy by Propensity score analysis.

BACKGROUND Acute subdural haematomas (ASDH) are found frequently following traumatic brain injury (TBI) and they are considered the most lethal type of mass lesions. The decision to perform a procedure to evacuate ASDH and the approach, either via craniotomy or decompressive craniectomy (DC), remains controversial. METHODS We reviewed a prospectively collected series of 343 moderate to severe TBI patients in whom ASDH was the main lesion (ASDH volumes ≥10cc). Patients with early comfort measures (early mortality prediction >50% and not ICP monitored), bilateral ASDH or the presence of another intracranial haematoma with volumes exceeding two times the volume of the ASDH were excluded. Among them, 112 were managed conservatively, 65 underwent ASDH evacuation by craniotomy and 166 by DC (103 pre-emptive DC, 63 obligatory DC). We calculated the average treatment effect by propensity score (PS) analysis using the following covariates: age, year, hypoxia, shock, pupils, major extracranial injury, motor score, MLS, ASDH volume, swelling, intraventricular and subarachnoid haemorrhage presence. Then, multivariable binary regression and ordinal logistic regression analysis were performed to estimate associations between predictors and mortality and 12 months-GOS respectively. The patients' inverse probability weights were included as an independent variable in both regression models. RESULTS The main variables associated with outcome were year, age, falls from patient´s own height, hypoxia, early deterioration, pupillary abnormalities, basal cistern effacement, compliance to ICP monitoring guidelines and type of surgical approach (craniotomy and preemptive DC). According to sliding dichotomy analysis, we found that patients in the intermediate or worst bands of unfavourable outcome prognosis seemed to achieve better than expected outcome if they underwent pre-emptive DC rather than craniotomy. CONCLUSIONS When differences in patient's baseline characteristics are balanced by PS, the variables associated with long-term outcome are year, age, falls from patient´s own height, hypoxia, early deterioration, pupillary abnormalities, basal cistern effacement, compliance to ICP monitoring guidelines and type of surgical approach (craniotomy and pre-emptive DC are associated with better outcome). Patients with an intermediate or worse risk of unfavourable outcome according to their baseline characteristics might achieve better than expected outcome if they undergo pre-emptive DC.