Abstract
OBJECTIVE: To investigate the factors influencing acute cerebral hemorrhage (ACH) secondary to nosocomial gastrointestinal hemorrhage (GIH) and construct a nomogram prediction model. METHODS: A total of 500 ACH patients admitted to our hospital from August 2022 to August 2024 were retrospectively analyzed and divided into a modeling group (350 cases) and a validation group (150 cases) in a 7:3 ratio. Patients in the modeling group were further divided into the GIH and non-GIH groups. Clinical data were collected, and multivariate logistic regression was used to analyze risk factors. A nomogram model was constructed using R software. The predictive performance was evaluated using the ROC curve, calibration curve, and decision curve analysis (DCA). RESULTS: Among 500 patients, 78 (15.6%) developed GIH. In the modeling group (350 cases), 56 (16.0%) had GIH. There were significant differences in age, history of coronary heart disease, history of alcohol consumption, NIHSS score, systolic blood pressure, and hemorrhage volume between groups (P<0.05). Logistic regression analysis identified these factors as independent risk factors for secondary GIH (P<0.05). The Area Under Curve(AUC) was 0.798 in the modeling group and 0.978 in the validation group, with calibration curves showing good agreement between predicted and observed values (Hosmer-Lemeshow(H-L) test: modeling group, χ²=7.156, P=0.732; validation group, χ²=7.015, P=0.703). DCA indicated a high clinical application value when the probability ranged from 0.06 to 0.95. CONCLUSION: Age, history of coronary heart disease, history of alcohol consumption, NIHSS score, systolic blood pressure, and hemorrhage volume are key risk factors for secondary GIH in ACH patients. The nomogram model constructed based on these factors demonstrates good predictive performance and clinical application value. It can help clinicians prevent early onset and reduce the risk of bleeding in patients.