Predictors of Outcome in Intracerebral Hemorrhage and Performance of Hemphill ICH Score and GVS Score

Background: The outcome of intracerebral hemorrhage (ICH) is gloomy. There are several scoring systems for predicting its outcome. Objective: The purpose of the present study was to observe the predictors of outcome in ICH patients and to assess the performance of ‘Hemphill ICH score’ and ‘GVS score’. Methodology: This cohort study involved patients of ICH admitted within 72 hours of acute event in July to December 2017 in the Department of Neurology at National Institute of Neurosciences (NINS) & Hospital, Dhaka, Bangladesh. Clinical and radiological data at admission and in-hospital events were obtained from medical records. Patients who were discharged from the hospital were interviewed at 30th day after event by face to face interview or over telephone. Follow up data was not found for 4 patients. Results: This study involved 115 patients of ICH [median age 60 years (interquartile range, IQR 50-70); 46% (53/115) female]. In-hospital and 30-day mortality of the ICH patients was 22% (25/115: 95% CI 15-30%) and 38% (42/111; 95% CI 29–48%) respectively. There was almost perfect agreement between Hemphill ICH score and GVS score (p<0.001 and κ=0.862). Kaplan-Meier survival curves of patients with different Hemphill ICH scores revealed significant difference among them; which was also true for patients with different GVS scores (by log-rank test, p<0.001 for both). The area under the curve (AUC) for the Hemphill ICH score to predict mortality was 0.70 (95% CI, 0.60–0.80) and for the GVS score 0.74 (95% CI, 0.64–0.83). In bivariate logistic regression, NIHSS score, GCS score, blood glucose at admission, nosocomial pneumonia, ICH volume, midline shift along with both Hemphill ICH and GVS score were observed to predict mortality at 30 days (p<0.05 for all). In two separate models adjusting for blood glucose and nosocomial pneumonia, every 1-point increase in the Hemphill ICH score and GVS score increases the mortality risk by 2.35 fold (adjusted OR: 2.35; 95% CI 1.33-4.16; p=0.003) and 2.99 fold (adjusted OR: 2.99; 95% CI 1.57-5.72; p=0.001) respectively. Conclusions: Both Hemphill ICH and GVS score have comparable predicting ability of outcome in ICH. In addition to components of scoring systems, occurrence of nosocomial pneumonia and blood glucose seems important. Journal of National Institute of Neurosciences Bangladesh, January 2021, Vol. 7, No. 1, pp. 3-9