The Intracerebral Hemorrhage Score: A Self-Fulfilling Prophecy?

Abstract BACKGROUND The ICH Score has become the standard for risk-stratification of 30-d mortality in patients with intracerebral hemorrhage (ICH), but treatment has evolved over the last 17 yr since its inception. We sought to determine if the ICH Score remains an accurate predictor of 30-d mortality in these high acuity patients. OBJECTIVE To determine the role the ICH Score has on mortality in current treatment of patients. METHODS A retrospective review of 554 patients treated for acute, spontaneous ICH at 2 large academic institutions between 2010 and 2014 was carried out. Surgical intervention in the form of external ventricular drain or craniotomy was performed when indicated. All patients were managed medically until discharge or death. RESULTS Over half (53.6%) of the patients presented with ICH of the basal ganglia/thalamus and the majority (71%) presented with ICH Scores of 0 to 2. Overall mortality was 25.1%. Observed mortality in moderate grade ICH Score patients (3 and 4) was lower than expected (49% vs 72%, P < .001) and (71% vs 97%, P < .001) when compared to the original ICH Score results. Despite differences in ICH and intraventricular hemorrhage volume, and Glasgow Coma Scale there was no difference in surgical intervention (12.2% vs 11.8%, P = .94) between the two groups. Withdrawal of care was instituted in 56.6% of all patients who died and increased with ICH Score. CONCLUSION In our cohort, the original ICH score did not accurately predict the mortality rate. Patient survival exceeded ICH Score-predicted mortality regardless of surgical intervention. Reevaluation of predictive scores could be useful to aid in more accurate prognoses.