Basal ganglia haemorrhages: efficacy and limits of different surgical strategies

To assess the role of surgery in patients with spontaneous basal ganglia haemorrhages, we evaluated poor outcome (mortality and prolonged unawareness) one month after 'open' surgery in patients with haematomas larger than 30 cm(3). One hundred and twenty-seven patients were traced over a 5-year period. Excluding deeply comatose patients (Glasgow Coma Scale [GCS] 3-4, n = 39), we analysed the remaining 88 patients, dividing them into two homogeneous groups according to the modality of treatment: aggressive or palliative. Multivariate analysis was applied both to the overall population and to the two groups in order to determine factors prognostic for poor outcome. Aggressive treatment was defined as surgery as the first-choice treatment modality aimed at 'complete' evacuation. Palliative treatment was defined as delayed surgery and/or surgery aimed at clot removal only to obtain internal decompression. Efficacy was assessed in patients having the same initial GCS score in both groups. Factors significantly associated with outcome were preoperative complications, volume, timing of operation, residual clots and postoperative complications. Outcomes were significantly better for aggressive surgery (17% vs. 68%, p < 0.001). On analysing the two treatment groups, volume and GCS were found to be significantly correlated with outcome in the palliative treatment group, while pre- and postoperative complications were significantly correlated with outcome in the aggressive treatment group. As judged by preoperative GCS score, aggressive treatment is always effective while palliative treatment is valid for GCS 9 or more. It clearly emerged that early surgery, aimed at removing all the clots, improves the outcome in patients with spontaneous ganglionic haemorrhages (excluding deeply comatose patients) and has wider indications than palliative surgery. This aggressive strategy is negatively affected by pre- and postoperative general complications.