Predicting mortality in spontaneous intracerebral hemorrhage: can modification to original score improve the prediction?

BACKGROUND AND PURPOSE A clinical grading scale for intracerebral hemorrhage (ICH), formally ICH score, was recently developed showing to predict 30-day mortality in a simple and reliable manner. The aim of the present study was to validate the original ICH (oICH) score in an independent cohort of patients from a developing country assessing 30-day mortality and 6-month functional outcome and whether its modifications can improve prediction. METHODS Consecutive patients admitted with acute ICH between January 1, 2003, and July 31, 2004, were prospectively included. oICH score was applied and 2 modified ICH (mICH) scores were created with the same variables, except localization, of the oICH score but with different cutoff values. Outcome was assessed as 30-day mortality and 6-month good outcome (Glasgow Outcome Scale [GOS] 4 to 5). RESULTS A total of 153 patients were included during study period. Thirty-day mortality rate was 34.6% (n=53), and 59 patients (38.6%) had good functional outcome (GOS 4 to 5) at 6 months. The oICH and mICH scores predicted mortality equally well. According to Youden's index (J), the oICH score was a reliable predictor for mortality (J=0.59) but less reliable for predicting good outcome (J=0.54). The mICH scores were equal in predicting mortality but better for predicting good outcome than the oICH score (J=0.60). CONCLUSIONS oICH score also confirms its validity in a socially and culturally different population. Modifications of oICH do not improve its 30-day mortality prediction but improve its ability to predict good functional outcome at 6 months.