A nomogram to predict the risk of postoperative intracranial rebleeding in patients with spontaneous intracranial hematoma.

Abstract

The 30-day mortality rate of spontaneous cerebral hemorrhage (ICH) is approximately 30-50%. Surgery may improve the prognosis of patients with severe ICH. However, ICH survivors after surgery still face the risks of postoperative intracranial rebleeding (PIB), and clinical tools that accurately predict the risk of PIB occurrence are not available. Therefore, a retrospective study was performed. The population was divided into two groups according to the occurrence of PIB. Univariate and multivariate logistic regression analyses were performed to screen risk factors for PIB. Next, an early PIB risk nomogram prediction model was constructed. In addition, the impact of PIB on the prognosis of ICH was evaluated. In total, 150 ICH patients were continuously enrolled in this study; 21 patients suffered from PIB, and the overall incidence of PIB was 14.0% (21/150). Coronary heart disease history, a lower GCS score, and subarachnoid hemorrhage absence were screened as risk factors for early PIB. The early PIB risk nomogram showed good calibration and discrimination with a concordance index of 0.807 (95% confidence interval (CI), 0.715-0.899), which was confirmed to be 0.788 through bootstrapping validation. In addition, a significant difference in discharged GOS scores (P = 0.043) was observed between the PIB group and the n-PIB group. These results showed that a history of coronary heart disease, a lower GCS score, and absence of subarachnoid hemorrhage were risk factors for early PIB. In addition, the early PIB risk nomogram prediction model exhibits good discrimination and calibration. The occurrence of PIB could reduce the prognosis of ICH patients.