A simple scoring system for accurate prediction of outcome within four days of a subarachnoid haemorrhage

This study was designed to examine the consistency of a number of easily identifiable predictive factors in assessing outcome within four days of a subarachnoid haemorrhage. Patients with a proven subarachnoid haemorrhage, aged between 15-65, of any neurological grade who had bled within 72 hours of admission, and who had undergone a CT scan within 96 hours of the ictus, were included. Three groups of patients were studied prospectively. The studies were separated in time and place. The series were similar overall but there were some variations between the three groups of patients because of alterations in referral patterns and management strategies between the series. There were significant differences in the patients' ages, grades on admission, timing of angiography, negative angiography rate and timing of operation. This did not affect overall outcome; 57%, 61% and 59% of the patients in series 1, 2 and 3 respectively making a good recovery. The proportion of patients with a poor outcome was also similar. To identify the level of risk of an individual patient within the first few days of haemorrhage, we considered a number of early predictive factors. Two emerged as strong predictors of outcome; the early neurological grade and the distribution of blood on the CT scan. We developed a simple scoring system from the first series, based on these findings, designed to predict outcome at three months. The scoring system was calculated on the basis of the distribution of blood seen on the CT scan and the patients' neurological grade on admission. Two points each were scored for interhemispheric, intraventricular, basal or intracerebral blood (excluding blood in the sylvian fissures). Patients in grade 1-3 scored -1, grade 4 scored 0, grades 5 & 6 scored +5. The scan score and grade score were summated to give the overall score. Patients were placed in risk groups (low, score -1; medium, score 0-2; high, score 3+). The scoring system was then applied prospectively to the two subsequent groups of patients. In each of the three series there was a clear correlation between the patients' scores and their outcomes but more importantly the probability of each outcome for each risk group was considered. In both the second and third series the probability of a full recovery in the low risk group was very likely--P = 0.000