Comments on the Risk Stratification for the In-Hospital Mortality in Subarachnoid Hemorrhage: The HAIR Score

Abstract

The HAIR score was published in the August 2014 issue ofNeurocritical Care as a tool to enable risk stratification ofin-hospital mortality after subarachnoid hemorrhage (SAH)[1]. The scale features four variables: Hunt and Hess score,age, intraventricular hemorrhage, and re-bleeding thatcomprise a composite maximum score of 8. Several otherfactors, previously described as being associated with pooroutcome after SAH, were not considered in the HAIRscore, for example larger aneurysm size or the presence ofglobal cerebral edema on follow-up CT scan. In their ret-rospectively reviewed cohort of 400 patients, Lee et al.demonstrated increasing mortality rates with incrementalincreases in HAIR scores across the entire spectrum ofscores between 0 and 7.We sought to externally validate the HAIR score in ourcohort of SAH patients from the Columbia University SAHOutcomes Project (SHOP) registry. Data from 1,629 pa-tients were prospectively collected and adjudicated weeklyby the study team. Our findings show that the HAIR scoreadequately discriminates mortality rates between scores 0and 5, however, fails to discriminate between scores 6 and7 (Fig. 1a). The area under the ROC-curve was 0.90(Fig. 1b).Additionally, Lee et al’s study did not include any pa-tients with a HAIR score of 8, which, in our cohort,corresponded to a 100 % in-hospital mortality rate.In summary, analysis of our larger, prospectively col-lected cohort of patients with SAH confirmed the authors’