This study aimed to identify independent risk factors for in-hospital complications among patients with spontaneous intracerebral hemorrhage (SICH) and to develop an admission-based prediction model to enable early recognition and timely intervention. We retrospectively analyzed SICH patients admitted to the neuro-intensive care unit (NICU) from June 2019 to June 2022. Patients with secondary hemorrhage or missing key data were excluded. The primary endpoint was any predefined in-hospital complication, including infection, rebleeding, or seizures. Admission demographics, vital signs, laboratory findings, imaging features, and perioperative data were collected. Data were randomly split into training and test datasets (7:3). Prediction models were built using multivariate logistic regression, Lasso, random forest, support vector machine, and decision tree approaches, with cross-validation to avoid overfitting. Model performance was assessed by the area under the receiver operating characteristic curve (AUC), Brier score, accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), calibration, and decision curve analysis. Among 224 patients, 129 (57.6%) developed at least one complication. The logistic regression model incorporating HbA1c, GCS score, diastolic blood pressure, basal ganglia involvement, and craniotomy achieved the best performance (AUC = 0.867; Brier = 0.176) with favorable clinical benefit on DCA. The proposed admission-based model demonstrated good discrimination, calibration, and clinical utility for predicting in-hospital complications in SICH. HbA1c and GCS were strong predictors linking impaired glucose metabolism and decreased consciousness to elevated complication risk. This simple, interpretable model may assist early risk assessment and optimize management of SICH patients in the NICU. This model is intended for risk prediction rather than causal inference.
[1]Fudan Univ, Dept Neurosurg & Neurocrit care, Huashan Hosp, Shanghai 200040, Peoples R China.
[2]Fudan Univ, Natl Ctr Neurol Disorders,Neurosurg Inst,Dept Neur, Shanghai Clin Med Ctr Neurosurg,Neural Regenerat, Shanghai Key Lab Brain Funct & Restorat & Neural R, Shanghai 200040, Peoples R China.
[3]Shanghai Jiao Tong Univ, Xinhua Hosp, Sch Med, Dept Neurosurg, Shanghai 200092, Peoples R China.
[4]Shanghai Jiao Tong Univ, Tongren Hosp, Dept Infect Dis, Sch Med, Shanghai 200336, Peoples R China.
[5]Fudan Univ, Huashan Hosp, Dept Neurosurg, 12 Wulumuqi Zhong Rd, Shanghai 200040, Peoples R China.
(8.0+极速模式)