A Nomogram Model for Predicting Prognosis in Spontaneous Intracerebral Hemorrhage Patients

Intracranial hemorrhage is the second most common stroke subtype following ischemic stroke and usually induces high mortality and disability. Here, we conducted a retrospective study to establish a nomogram clinical prediction model.

First, the baseline data of patients who presented to our hospital in 2015-2021 were collected and compared (789 patients for the training cohort and 378 patients for the validation cohort). Second, univariate and binary logistic analyses were performed to screen out alternative indicators. Finally, a clinical prediction model by nomogram was established that included such indicators to estimate the prognosis of intracranial hemorrhage patients.

Univariate logistic analysis was used to screen several possible impact factors, including hypertension, hematoma volume, Glasgow Coma Scale (GCS) score, intracranial hemorrhage (ICH) score, irregular shape, uneven density, intraventricular hemorrhage (IVH) relation, fibrinogen, D-dimer, low density lipoprotein (LDL), high-density lipoprotein (HDL), creatinine, total protein, hemoglobin (HB), white blood cell (WBC), neutrophil blood cell (NBC), lymphocyte blood cell (LBC), the neutrophil lymphocyte ratio (NLR), surgery, deep venous thrombosis (DVT) or pulmonary embolism (PE) rate, hospital day, and hypertension control. Further binary logistic analysis revealed that ICH score (p = 0.036), GCS score (p = 0.000), irregular shape (p = 0.000), uneven density (p = 0.002), IVH relation (p = 0.014), surgery (p = 0.000) were independent indicators to construct a nomogram clinical prediction model. The C statistic was 0.840.

ICH score, GCS score, irregular shape, uneven density, IVH relation, surgery are easily available indicators to assist neurologists in formulating the most appropriate therapy for every intracranial hemorrhage patient. Further large prospective clinical trials are needed to obtain more integrated and reliable conclusions.