Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 430030 Wuhan, Hubei, China.
J Integr Neurosci. 2023 Feb 20;22(2):42. doi: 10.31083/j.jin2202042.
Intracranial hemorrhage is the second most common stroke subtype following ischemic stroke and usually induces high mortality and disability. Here, we conducted a retrospective study to establish a nomogram clinical prediction model.
First, the baseline data of patients who presented to our hospital in 2015-2021 were collected and compared (789 patients for the training cohort and 378 patients for the validation cohort). Second, univariate and binary logistic analyses were performed to screen out alternative indicators. Finally, a clinical prediction model by nomogram was established that included such indicators to estimate the prognosis of intracranial hemorrhage patients.
Univariate logistic analysis was used to screen several possible impact factors, including hypertension, hematoma volume, Glasgow Coma Scale (GCS) score, intracranial hemorrhage (ICH) score, irregular shape, uneven density, intraventricular hemorrhage (IVH) relation, fibrinogen, D-dimer, low density lipoprotein (LDL), high-density lipoprotein (HDL), creatinine, total protein, hemoglobin (HB), white blood cell (WBC), neutrophil blood cell (NBC), lymphocyte blood cell (LBC), the neutrophil lymphocyte ratio (NLR), surgery, deep venous thrombosis (DVT) or pulmonary embolism (PE) rate, hospital day, and hypertension control. Further binary logistic analysis revealed that ICH score ( = 0.036), GCS score ( = 0.000), irregular shape ( = 0.000), uneven density ( = 0.002), IVH relation ( = 0.014), surgery ( 0.000) were independent indicators to construct a nomogram clinical prediction model. The C statistic was 0.840.
ICH score, GCS score, irregular shape, uneven density, IVH relation, surgery are easily available indicators to assist neurologists in formulating the most appropriate therapy for every intracranial hemorrhage patient. Further large prospective clinical trials are needed to obtain more integrated and reliable conclusions.
脑出血是继缺血性脑卒中之后第二常见的脑卒中亚型，通常导致高死亡率和高残疾率。在这里，我们进行了一项回顾性研究，以建立一个列线图临床预测模型。
首先，收集并比较了 2015 年至 2021 年在我院就诊的患者的基线数据（训练队列 789 例，验证队列 378 例）。其次，进行单变量和二元逻辑分析以筛选出替代指标。最后，通过列线图建立了一个临床预测模型，该模型包含了这些指标，以估计脑出血患者的预后。
单变量逻辑分析筛选出了几个可能的影响因素，包括高血压、血肿量、格拉斯哥昏迷量表（GCS）评分、脑出血（ICH）评分、不规则形状、不均匀密度、脑室内出血（IVH）关系、纤维蛋白原、D-二聚体、低密度脂蛋白（LDL）、高密度脂蛋白（HDL）、肌酐、总蛋白、血红蛋白（HB）、白细胞（WBC）、中性粒细胞白细胞（NBC）、淋巴细胞白细胞（LBC）、中性粒细胞淋巴细胞比值（NLR）、手术、深静脉血栓形成（DVT）或肺栓塞（PE）发生率、住院天数和高血压控制情况。进一步的二元逻辑分析表明，ICH 评分（ = 0.036）、GCS 评分（ = 0.000）、不规则形状（ = 0.000）、不均匀密度（ = 0.002）、IVH 关系（ = 0.014）、手术（ = 0.000）是构建列线图临床预测模型的独立指标。C 统计量为 0.840。
ICH 评分、GCS 评分、不规则形状、不均匀密度、IVH 关系、手术是易于获得的指标，可以帮助神经科医生为每个脑出血患者制定最合适的治疗方案。需要进一步进行大型前瞻性临床试验以获得更综合和可靠的结论。
