Abstract W P321: Predicting Mortality in ICH: Redefining the ICH Score
Introduction: The stated 30-day mortality rate for the original ICH score is 0%, 13%, 26%, 72%, 97%, and 99% for scores of 0-5, respectively. Although the score has been externally validated and widely used, recently proposed updates to the model have failed to improve its mortality prognostication. We hypothesized that precise quantification of ICH and IVH volumes more accurately reflect the severity of the disease and enhance its prognostic value.
Materials and Methods: A modified ICH score was developed to quantify ICH volume, IVH volume (with the modified Graeb score), GCS at admission, and hemorrhage location. The modified score was as follows: GCS 11-15: 0 points, 6-10: 1 point, 3-5: 2 points; ICH volume <20 cm3: 0 points, 20-34 cm3: 1 point, 35-44 cm3: 2 points, >45 cm3: 3 points; mGraeb score 0: 0 points, 1-9: 1 point, >10: 2 points; cerebellar hemorrhage: 1 point; brainstem hemorrhage: 1 point. Possible scores ranged from 0-9 and were designed to provide a more accurate stratified prognostic reflection for 30-day mortality.
Results: The modified score was tested in 362 patients previously evaluated with the original ICH score. Thirty -day mortality rates per group were 0%, 5%, 33%, 61%, 70%, and 91%, for ICH scores of 0-5, respectively. Median time to death in those that died within 30 days was 15.5, 3, 4, 1.5, and 1 day(s) for ICH scores of 1-5, respectively. Application of the proposed modified ICH score resulted in 30-day mortality rates of 0: 0%, 1: 5%, 2: 11%, 3: 4%, 4: 51%, 5: 73%, 6: 74%, 7: 90%, 8/9: 100%. Median time to death was 6, 13, 4, 4, 2, 2, 1, and 0 day(s), in scores of 1-8, respectively. In our patient population, 99% of patients who died within a 30-day period did so within 18 days. Furthermore, 75% died within the first week and 50% within the first 72 hours. Predicted 30-day mortality in the original ICH score had an area under the receiver-operating curve (AUROC) of 0.88, while the proposed new model showed a higher AUROC of 0.91.
Conclusions: Improved outcomes in ICH therapy require better assessment and prognostic tools. By accurately determining hemorrhage volumes and increasing anatomical localization, stratification of ICH patients can result in better predictors of 30-day mortality. Further validation of this model is needed in larger patient populations.
Author Disclosures: D.J. McCracken: None. B.P. Lovasik: None. M.E. McDougal: None. C.E. McCracken: None. G. Pradilla: None.
- © 2015 by American Heart Association, Inc.