Background and Objective: Spontaneous intracerebral haemorrhage (ICH) is more common in Asia than in western countries, and has a high mortality rate. A simple prognostic score for predicting grave prognosis of ICH is lacking. Our objective was to develop a simple and reliable score for most physicians.
Material and Method: ICH patients from seven Asian countries were enrolled between May 2000 and April 2002 for a prospective study. Clinical features such as headache and vomiting, vascular risk factors, Glasgow coma scale (GCS), body temperature (BT), blood pressure on arrival, location and size of haematoma, intraventricular haemorrhage (IVH), hydrocephalus, need for surgical treatment, medical treatment, length of hospital stay and other complications were analyzed to determine the outcome using a modified Rankin scale (MRS). Grave prognosis (defined as MRS of 5-6) was judged on the discharge date.
Results: 995 patients, mean age 59.5+14.3 years were analyzed, after exclusion of incomplete data in 87 patients. 402 patients (40.4%) were in the grave prognosis group (MRS 5-6). Univariable analysis and then multivariable analysis showed only four statistically significant predictors for grave outcome of ICH. They were fever (BT> 37.8oc), low GCS, large haematoma and IVH. The grave prognosis on spontaneous intracerebral haemorrhage (GP on STAGE) score was derived from these four factors using a multiple logistic model.
Conclusion: A simple and pragmatic prognostic score for ICH outcome has been developed with high sensitivity (82%) and specificity (82%). Furthermore, it can be administered by most general practitioners. Validation in other populations is now required.

Keywords: Cerebral haemorrhage, Stroke, Cerebrovascular disease, Intracerebral haemorrhage, Prognostic cerebral haemorrhage score