E-ISSN: 1019-5157
ISSN: 2651-5024
Research
Predicting Postoperative Rebleeding After Emergency Craniotomy in TBI: Development and Validation of a Comprehensive Nomogram
Jiahao Wang✉ ,
Haixiao Jiang ,
jingming Xu ,
Wei Zeng
Haixiao Jiang ,
jingming Xu ,
Wei Zeng
Neurosurgery, Yangzhou University
DOI: 10.5137/1019-5149.JTN.49932-25.3
Article in Press
Corresponding Author:
Jiahao Wang (1620355664@qq.com)
Abstract
Aim
Postoperative bleeding after emergency craniotomy for traumatic brain injury (TBI) continues to be a leading cause of death and long-term disability. Current predictive models of risk factors are based on a very limited number of variables or only cover small cohorts, thus limiting their usefulness in the clinical setting. The objective of this study was to create a robust multidimensional model that could define patients at high risk of rebleeding and manage them more appropriately.
Material and Methods
A total of 540 cases of TBI receiving emergency craniotomy were included and were randomly divided into two groups (training group (n=360) and validation group (n=180)). Clinical information obtained within 24 hours of admission included sex, age, and neurological assessment, while imaging data (such as hematoma volume), laboratory tests (such as coagulation profile) and intraoperative data (such as blood pressure fluctuations) were also collected. We implemented the logistic regression analysis to identify independent predictors and then generated a nomogram. Concordance index and decision curve analysis (DCA) were also used to assess this nomogram.
Results
The overall percentage of patients who had postoperative bleeding was found to be 23.7%. Independent predictors included prior anticoagulant/antiplatelet use and cerebrovascular disease, larger hematoma volume, older age, diabetes, coagulation abnormalities (elevated INR, prolonged APTT), increased number of small contusions, intraoperative acute brain bulging, and intraoperative hypotension. The nomogram demonstrated good discrimination (C-index: 0.880 [95% CI: 0.8240.897] in the training group; 0.807 [95% CI: 0.7800.889] in the validation group) and calibration. DCA indicated that the model provided clinical net benefit across a 1787% risk threshold, outperforming strategies of treating all or no patients.
Conclusion
This comprehensive nomogram could accurately classify TBI patients who may be at increased risk of developing postoperative bleeding and may serve as a reliable guide for perioperative management of TBI patients to optimize clinical outcomes.
Postoperative bleeding after emergency craniotomy for traumatic brain injury (TBI) continues to be a leading cause of death and long-term disability. Current predictive models of risk factors are based on a very limited number of variables or only cover small cohorts, thus limiting their usefulness in the clinical setting. The objective of this study was to create a robust multidimensional model that could define patients at high risk of rebleeding and manage them more appropriately.
Material and Methods
A total of 540 cases of TBI receiving emergency craniotomy were included and were randomly divided into two groups (training group (n=360) and validation group (n=180)). Clinical information obtained within 24 hours of admission included sex, age, and neurological assessment, while imaging data (such as hematoma volume), laboratory tests (such as coagulation profile) and intraoperative data (such as blood pressure fluctuations) were also collected. We implemented the logistic regression analysis to identify independent predictors and then generated a nomogram. Concordance index and decision curve analysis (DCA) were also used to assess this nomogram.
Results
The overall percentage of patients who had postoperative bleeding was found to be 23.7%. Independent predictors included prior anticoagulant/antiplatelet use and cerebrovascular disease, larger hematoma volume, older age, diabetes, coagulation abnormalities (elevated INR, prolonged APTT), increased number of small contusions, intraoperative acute brain bulging, and intraoperative hypotension. The nomogram demonstrated good discrimination (C-index: 0.880 [95% CI: 0.8240.897] in the training group; 0.807 [95% CI: 0.7800.889] in the validation group) and calibration. DCA indicated that the model provided clinical net benefit across a 1787% risk threshold, outperforming strategies of treating all or no patients.
Conclusion
This comprehensive nomogram could accurately classify TBI patients who may be at increased risk of developing postoperative bleeding and may serve as a reliable guide for perioperative management of TBI patients to optimize clinical outcomes.
Keywords
Traumatic brain injury
Postoperative rebleeding
Nomogram
Risk prediction
Emergency craniotomy