Surgical intervention in supratentorial intracerebral haemorrhage (ICH) is still controversial. We assessed the value of haematoma evacuation with a case-control study. 145 consecutive patients with supratentorial spontaneous ICH without tumour or vascular abnormalities were analysed. Haematoma evacuation was performed in 24 patients. Age, sex, Glasgow Coma Scale (GCS), level of consciousness, pupillary reaction on admission, localisation, aetiology and volume of the haematoma, presence of ventricular blood, and Glasgow Outcome Scale (GOS) on discharge were analysed. From statistical analysis 40 patients > 80 years and with haematoma volume < 10 ml, who were always treated conservatively, were excluded. Prognostic factors retained from a multiple regression model with the dichotomised GOS scale (GOS 1-3, 4 + 5) as response variable were GCS, haematoma volume and location. The only difference between all medically treated and "operated" patients was haematoma volume, which was larger in the "operated" patients. All 24 evacuated cases could be matched to a medically treated control regarding age, haematoma volume and location, GCS, and pupillary reaction. Significant differences between the two groups could not be detected. Outcome was not different between the two groups. After separating the sample into patients with and without ventricular haemorrhage, there was no different outcome between the two groups either. We conclude that haematoma evacuation did not improve outcome in supratentorial spontaneous ICH. Since haematomas were evacuated mainly in clinically deteriorating patients, our data suggest that the only effect of haematoma evacuation is to stop progressive deterioration rather than to improve overall clinical outcome.