高血压脑出血患者神经内镜颅内血肿清除术后大脑环池受压和形态分级与预后的关系

Relationship between compression and shape changes of the cerebral cistern ambiens in hypertensive cerebral hemorrhage patients after neural endoscopic intracranial hematoma evacuation and prognosis

摘要:

目的 探究高血压脑出血（HCH）患者神经内镜颅内血肿清除术（NEIHE）后大脑环池受压程度和形态分级与预后的关系。方法 选择2020年1月－2023年6月于该院接受NEIHE治疗的HCH患者246例。根据术后6个月格拉斯哥预后评分（GOS），将患者分为预后良好组（n = 158）和预后不良组（n = 88）。采用分层回归模型，分析临床病理特征与大脑环池受压程度和形态分级的关系。采用多因素Logistic回归模型，分析影响HCH患者NEIHE后预后不良的独立危险因素，并构建预测模型。以预后不良为结局变量，绘制受试者操作特征曲线（ROC curve）。比较不同预测模型的曲线下面积（AUC）、综合判别改善指数（IDI）和净重新分类指数（NRI）。采用Hosmer-Lemeshow拟合优度检验判断模型的拟合优度。采用广义相加模型（GAM），分析GOS与相关影响因素的关系。结果 与预后良好组相比，预后不良组术前出血量更多，术前格拉斯哥昏迷评分（GCS）更低，中线偏移、高血压病史 ≥ 10年、出血破入脑室、手术时机在晚期的患者占比更高，差异均有统计学意义（P < 0.05）。与预后良好组相比，预后不良组血肿同侧及对侧大脑环池受压3至4分和大脑环池形态分级为Ⅴ级的患者占比更高，差异均有统计学意义（P < 0.05）。分层回归分析显示，出血部位、中线偏移和颅内压对大脑环池受压程度及形态分级有影响，差异均有统计学意义（P < 0.05）。多因素Logistic回归分析（模型2：包含大脑环池受压程度和形态分级）结果显示，中线偏移、高血压病史 ≥ 10年、术前出血量 > 37 mL、出血破入脑室、术前GCS < 7分、手术时机晚期、大脑环池受压程度（血肿同侧3至4分，血肿对侧3至4分）和形态分级Ⅴ级，是影响HCH患者NEIHE后预后不良的独立危险因素（P < 0.05）。模型1（不包含大脑环池受压程度和形态分级）加入大脑环池受压程度和形态分级后，AUC提高到0.812（95%CI：0.763～0.872），且模型的IDI（0.081，95%CI：0.049～0.095）和NRI（0.611，95%CI：0.510～0.674）提高，差异均有统计学意义（P < 0.05）。与模型1比较，模型2的AUC（0.826，95%CI：0.771～0.863）、IDI（0.085，95%CI：0.052～0.110）和NRI（0.628，95%CI：0.510～0.709）提高，差异均有统计学意义（P < 0.05）。Hosmer-Lemeshow检验显示，预后不良预测模型2（P = 0.878）较模型1（P = 0.691）的拟合优度更理想。GAM分析显示，大脑环池受压程度越严重和形态分级越高，GOS评分越低，差异有统计学意义（P < 0.05）。ROC curve分析显示，大脑环池受压程度和形态分级联合应用，对预后不良的预测价值较高，AUC为0.935（95%CI：0.890～0.971），敏感度为70.26%，特异度为93.84%。结论 HCH患者NEIHE后大脑环池受压程度和形态改变与预后密切相关，且联合应用对预后不良具有一定的预测价值。出血部位、中线偏移和颅内压等因素，明显影响大脑环池受压程度和形态改变。

Abstract:

Objective To explore the relationship between the degree of compression and shape grading of the cerebral cistern ambiens in hypertensive cerebral hemorrhage (HCH) patients after neural endoscopic intracranial hematoma evacuation (NEIHE) and prognosis.Methods 246 HCH patients who underwent NEIHE surgery from January 2020 to June 2023 were selected as the research subjects. According to the postoperative 6 month Glasgow outcome score (GOS), the patients were divided into good prognosis group (n = 158) and poor prognosis group (n = 88). Hierarchical regression model was used to analyze the relationship between clinical pathological features and the degree of compression and shape grading of the cerebral cistern ambiens. Multivariate logistic regression model was used to analyze the independent risk factors affecting the poor prognosis of patients with HCH after NEIHE, and a prediction model was constructed. Receiver operating characteristic curve (ROC curve) was drawn with poor prognosis as the outcome variable. Area under the curve (AUC), integrated discrimination improvement (IDI) and net reclassification improvement (NRI) of different prediction models were compared. Hosmer-Lemeshow was used to test the goodness of fit. Generalized additive model (GAM) was used to analyze the relationship between GOS and related factors.Results Compared with the good prognosis group, the poor prognosis group had more preoperative blood loss, lower preoperative Glasgow coma score (GCS), and higher proportion of patients with midline deviation, hypertension history ≥ 10 years, hemorrhage breaking into the brain ventricle, and operation time in the late stage, the differences were statistically significant (P < 0.05). Compared with the good prognosis group, the proportions of patients with ipsilateral and contralateral cerebral cistern ambiens compression of 3 to 4 points and the morphological classification of cerebral cistern ambiens of grade V in the poor prognosis group were higher, and the differences were statistically significant (P < 0.05). The location of bleeding, midline deviation, and intracranial pressure all had an impact on the degree of compression and morphological grading of the cerebral cistern ambiens, and the differences were statistically significant (P < 0.05). Multivariate Logistic regression analysis (model 2: including the degree of compression and shape grading of the cerebral cistern ambiens) showed that midline deviation, history of hypertension ≥10 years, preoperative bleeding volume > 37 mL, bleeding into the brain ventricles, preoperative GCS < 7 points, late surgical timing, degree of compression of the cerebral cistern ambiens (3 to 4 points on the same side of hematoma, 3 to 4 points on the opposite side of hematoma), and grade V morphology were all risk factors influencing the poor prognosis of patients with HCH after NEIHE (P < 0.05). After incorporating the degree of compression and morphological grading of the cerebral cistern ambiens into model 1 (excluding the degree of compression and shape grading of the cerebral cistern ambiens), the AUC increased to 0.812 (95%CI: 0.763 ~ 0.872), and both the IDI (0.081, 95%CI: 0.049 ~ 0.095) and NRI (0.611, 95%CI: 0.510 ~ 0.674) of the model improved, with statistically significant differences (P < 0.05). Compared with model 1, model 2 showed an increase in AUC (0.826, 95%CI: 0.771 ~ 0.863), IDI (0.085, 95%CI: 0.052 ~ 0.110), and NRI (0.628, 95%CI: 0.510 ~ 0.709), with statistically significant differences (P < 0.05). The Hosmer-Lemeshow test showed that model 2 (P = 0.878) had a better fit for predicting poor prognosis than model 1 (P = 0.691). GAM analysis showed that the higher the degree of compression and morphological grading of the cerebral cistern ambiens, the lower the GOS, and the difference was statistically significant (P < 0.05). ROC curve analysis showed that the combined application of degree of compression and morphological grading of the cerebral cistern ambiens had a high predictive value for poor prognosis, with AUC of 0.935 (95%CI: 0.890 ~ 0.971), sensitivity of 70.26%, and specificity of 93.84%.Conclusion The degree of compression and shape changes of the cerebral cistern ambiens in HCH patients after NEIHE are closely related to prognosis, and the combined application has certain predictive value for poor prognosis. The location of bleeding, midline deviation, intracranial pressure, and other factors significantly affect the compression degree and shape change of the cerebral cistern ambiens.