Constructing and Validating a Nomogram for Survival in Patients without Hypertension in Hypertensive Intracerebral Hemorrhage-Related Locations

Gui-Jun Zhang,Hao Wang,Li-Chuan Gao,Jie-Yi Zhao,Tao Zhang,Chao You,Xiao-Yu Wang

DOI: https://doi.org/10.1016/j.wneu.2023.01.006

Abstract:Objective: We aimed to evaluate the risk factors for patients, who had hypertensive intracerebral hemorrhage (ICH)-specific location hemorrhage without hypertensive history, to elucidate a novel and detailed understanding. Methods: We conducted a retrospective review to identify patients diagnosed with hemorrhage in hypertensive ICH-specific locations without hypertensive history between January 2011 and December 2019 from West China Hospital. A least absolute shrinkage and selector operation (LASSO) algorithm was used to select the optimal prognostic factors, and then we performed a multivariable logistic analysis. To verify the accuracy of the nomogram in predicting patient outcome, we used Harrell's statistics, area under the curve, and a calibration as well as decision curves. Results: The LASSO method, at a tenfold cross-validation for 7-day mortality, 90-day mortality, and 90-day morbidity, was applied to construct the prognosis-predicting models. Both a higher Glasgow Coma Scale (GCS) score at admission and larger hematoma volume ≥13.64 mL were independently associated with better survival at 7 days and 90 days in multivariate analysis. Lactic dehydrogenase >250 IU/L and neutrophilic granulocyte/lymphocyte ratio in 1 increase were significantly associated with poor outcome at 90 days. Only one factor (GCS score at 7 days) influencing 90-day morbidity remained in a LASSO model. Conclusions: In this study, the GCS score, hematoma volume, and other laboratory factors (Lactic dehydrogenase and neutrophilic granulocyte/lymphocyte ratio) were related to survival. Our current findings of the specific location ICH need to be proven by a large randomized controlled trial study.