Abstract W P323: Development of a Nomogram to Predict Hematoma Expansion in Patients with Spontaneous Intracerebral Hemorrhage
Background: Early identification of intracerebral hemorrhage (ICH) patients at risk of significant hematoma expansion (SHE) could facilitate the selection of appropriate patients who are likely to benefit from therapies aiming to minimize ICH growth. Nomograms have been proved to have superior individualized disease-related risk estimations of given outcomes. This study aims to develop a normogram that can be performed during the hyperacute phase to predict the risk of SHE in patients with spontaneous ICH.
Methods: We reviewed clinical, laboratory, and radiological data from 237 patients diagnosed with spontaneous ICH who had baseline head CT within 12 hours of symptom onset and follow-up CT during the following 72 hours. SHE was defined as an absolute increase in ICH volume > 6ml or an increase greater than 33% from baseline to follow-up CT. To construct the nomogram, we performed logistic regression analyses to determine the predictors of SHE. Each predictor was assigned a point in the graphic interface of a nomogram, and the points were summed up to determine the predicted probability of SHE for a specific ICH patient.
Results: SHE occurred in 74 patients (31.2%). The final model to predict SHE, presented as a nomogram, included: time from onset to baseline CT scan (< 3h vs 3-12h), dementia, current smoking, antiplatelet use, serum creatinine level, Glasgow Comma Scale score, and presence of subarachnoid hemorrhage on baseline CT. The model had satisfactory discrimination ability with a bootstrap corrected c index of 0.77 (95% CI, 0.75-0.82) and good calibration. The in-hospital mortality was higher in patients with SHE (42% vs. 15%; p <0.001).
Conclusion: We developed and internally validated a novel nomogram model which accurately predicts the possibility of SHE based on seven easily obtainable parameters. This could be useful for treatment decision and stratification. External validation of our nomogram is warranted before its application to other populations.
Author Disclosures: X. Yao: None. M. Selim: None. Y. Xu: None. E. Siwila-Sackman: None.
- © 2015 by American Heart Association, Inc.