Objective:To explore the clinical characteristics of hydrocephalus patients with delayed intracranial hemorrhage (DICH) after ventriculo-peritoneal shunt (VPS) and independent influencing factors for DICH after VPS, and construct a nomogram model for predicting the incidence of DICH after VPS.Methods:A retrospective analysis was performed; clinical data of 168 patients with hydrocephalus who underwent VPS at Department of Neurosurgery, Lihuili Hospital Affiliated to Ningbo University from January 2016 to December 2023 were collected. Based on the postoperative DICH occurrence, these patients were categorized into a non-DICH group and a DICH group. Clinical characteristics of patients from the DICH group were analyzed. Differences in clinical data between patients from the non-DICH group and DICH group were compared. Multivariate Logistic regression analysis was performed to screen the independent influencing factors for DICH after VPS, and a nomogram model for predicting DICH incidence after VPS was constructed. Receiver operating characteristic (ROC) curve, calibration curve, Hosmer-Lemeshow test and Bootstrap resampling method were used to evaluate the predictive efficiency and stability of the nomogram model.Results:(1) Among the 168 patients, 36 (21.43%) developed DICH, which occurred (6.06±2.77) days after VPS (ranged 2-13 days). Seven patients presented with symptoms as vomiting, epilepsy, and decreased consciousness, while 29 patients had no hemorrhage-related complication. The hemorrhage volume in these 36 DICH patients was 0.85(0.30, 3.54) mL. Regarding the hemorrhage sites, 12 patients had simple puncture tract hemorrhage, 14 had simple intraventricular hemorrhage, 7 had both puncture tract hemorrhage and intraventricular hemorrhage, and 3 had puncture tract hemorrhage combined with hemorrhage in other sites. At discharge, 8 patients had a good prognosis, and 28 had a poor prognosis. (2) Compared with the non-DICH group, the DICH group had statistically older age, higher proportion of patients with hypertension or craniotomy history, lower preoperative neutrophil-to-lymphocyte ratio (NLR), higher postoperative NLR, and higher ratio of postoperative to preoperative NLR (NLRR), with significant differences ( P<0.05). Results of multivariate Logistic regression analysis indicated that age ( OR=1.077, 95% CI: 1.022-1.135, P=0.005), history of craniotomy ( OR=3.547, 95% CI: 1.384-9.092, P=0.008), and NLRR ( OR=1.744, 95% CI: 1.179-2.581, P=0.005) were independent influencing factors for DICH in hydrocephalus patients after VPS ( P<0.05). (3) A nomogram model for predicting DICH in hydrocephalus patients after VPS was constructed based on age, history of craniotomy and NLRR. ROC curve analysis showed that area under the curve (AUC) of the nomogram model for predicting DICH in hydrocephalus patients after VPS was 0.812 (95% CI: 0.735-0.890, P<0.001); Hosmer-Lemeshow test indicated that the model had good calibration, and no significant difference was noted between the predicted probabilities and actual probabilities ( χ2=10.487, P=0.233); decision curve analysis results showed that the nomogram model could generate good clinical net benefits with risk threshold of 0%-60%; AUC of the internal validation ROC curve was 0.805 (95% CI: 0.785-0.816) and the C-index was 0.82 by Bootstrap resampling method, suggesting good stability of the model. Conclusion:Hydrocephalus patients with advanced age, history of craniotomy, and elevated NLRR trend to have DICH following VPS; the nomogram model constructed based on age, history of craniotomy and NLRR has good predictive performance and clinical practicability.

Objective:To explore the clinical characteristics of hydrocephalus patients with delayed intracranial hemorrhage (DICH) after ventriculo-peritoneal shunt (VPS) and independent influencing factors for DICH after VPS, and construct a nomogram model for predicting the incidence of DICH after VPS.Methods:A retrospective analysis was performed; clinical data of 168 patients with hydrocephalus who underwent VPS at Department of Neurosurgery, Lihuili Hospital Affiliated to Ningbo University from January 2016 to December 2023 were collected. Based on the postoperative DICH occurrence, these patients were categorized into a non-DICH group and a DICH group. Clinical characteristics of patients from the DICH group were analyzed. Differences in clinical data between patients from the non-DICH group and DICH group were compared. Multivariate Logistic regression analysis was performed to screen the independent influencing factors for DICH after VPS, and a nomogram model for predicting DICH incidence after VPS was constructed. Receiver operating characteristic (ROC) curve, calibration curve, Hosmer-Lemeshow test and Bootstrap resampling method were used to evaluate the predictive efficiency and stability of the nomogram model.Results:(1) Among the 168 patients, 36 (21.43%) developed DICH, which occurred (6.06±2.77) days after VPS (ranged 2-13 days). Seven patients presented with symptoms as vomiting, epilepsy, and decreased consciousness, while 29 patients had no hemorrhage-related complication. The hemorrhage volume in these 36 DICH patients was 0.85(0.30, 3.54) mL. Regarding the hemorrhage sites, 12 patients had simple puncture tract hemorrhage, 14 had simple intraventricular hemorrhage, 7 had both puncture tract hemorrhage and intraventricular hemorrhage, and 3 had puncture tract hemorrhage combined with hemorrhage in other sites. At discharge, 8 patients had a good prognosis, and 28 had a poor prognosis. (2) Compared with the non-DICH group, the DICH group had statistically older age, higher proportion of patients with hypertension or craniotomy history, lower preoperative neutrophil-to-lymphocyte ratio (NLR), higher postoperative NLR, and higher ratio of postoperative to preoperative NLR (NLRR), with significant differences ( P<0.05). Results of multivariate Logistic regression analysis indicated that age ( OR=1.077, 95% CI: 1.022-1.135, P=0.005), history of craniotomy ( OR=3.547, 95% CI: 1.384-9.092, P=0.008), and NLRR ( OR=1.744, 95% CI: 1.179-2.581, P=0.005) were independent influencing factors for DICH in hydrocephalus patients after VPS ( P<0.05). (3) A nomogram model for predicting DICH in hydrocephalus patients after VPS was constructed based on age, history of craniotomy and NLRR. ROC curve analysis showed that area under the curve (AUC) of the nomogram model for predicting DICH in hydrocephalus patients after VPS was 0.812 (95% CI: 0.735-0.890, P<0.001); Hosmer-Lemeshow test indicated that the model had good calibration, and no significant difference was noted between the predicted probabilities and actual probabilities ( χ2=10.487, P=0.233); decision curve analysis results showed that the nomogram model could generate good clinical net benefits with risk threshold of 0%-60%; AUC of the internal validation ROC curve was 0.805 (95% CI: 0.785-0.816) and the C-index was 0.82 by Bootstrap resampling method, suggesting good stability of the model. Conclusion:Hydrocephalus patients with advanced age, history of craniotomy, and elevated NLRR trend to have DICH following VPS; the nomogram model constructed based on age, history of craniotomy and NLRR has good predictive performance and clinical practicability.