临床特征联合炎症标志物对机械取栓后急性缺血性脑卒中患者预后的预测

周玲玲 孟旭晨 钟伟杰 孙兆良 施小红 邓谭君 梅子贤 肖杰曦 汤定中 李轶

引用本文: 周玲玲,孟旭晨,钟伟杰,等. 临床特征联合炎症标志物对机械取栓后急性缺血性脑卒中患者预后的预测[J]. 海军军医大学学报,2025,46(10):1290-1296.DOI: 10.16781/j.CN31-2187/R.20250086.
Citation: ZHOU L, MENG X, ZHONG W, et al. Clinical characteristics combined with inflammatory markers for predicting prognosis of patients with acute ischemic stroke after mechanical thrombectomy[J]. Acad J Naval Med Univ, 2025, 46 (10): 1290-1296. DOI: 10.16781/j.CN31-2187/R.20250086.

临床特征联合炎症标志物对机械取栓后急性缺血性脑卒中患者预后的预测

doi: 10.16781/j.CN31-2187/R.20250086
基金项目: 

上海交通大学医学院附属第九人民医院“交叉”研究基金 JYJC202131.

详细信息

Clinical characteristics combined with inflammatory markers for predicting prognosis of patients with acute ischemic stroke after mechanical thrombectomy

Funds: 

"Cross" Research Fund of Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine JYJC202131.

  • 摘要:  目的 探索急性缺血性脑卒中(AIS)患者接受血管内机械取栓治疗后的潜在预后因素,并构建一个有效的预测模型。 方法 回顾性收集了来自2家卒中中心的202例前循环大血管闭塞性AIS患者的临床数据。所有患者均接受了血管内机械取栓治疗,治疗和随访至少持续90 d。收集患者的基本人口学特征、病历信息及基线血液生物标志物数据,通过最小绝对收缩和选择算子(LASSO)-logistic回归分析筛选出90 d后AIS的潜在预后指标。 结果 饮酒史(P=0.029)、高血压(P=0.001)、糖尿病(P=0.021)、卒中或短暂性脑缺血发作史(P=0.049)、入院时收缩压(P=0.009)、入院时舒张压(P=0.038)、血糖(P=0.003)、白细胞计数(P=0.001)、中性粒细胞计数(P=0.001)、纤维蛋白原(P=0.010)、系统性免疫炎症指数(P=0.008)及中性粒细胞与淋巴细胞比值(NLR)(P<0.001)与AIS不良临床结局相关。通过LASSO-logistic回归筛选出9个显著的预后决定因素。多因素logistic回归分析显示,男性(P=0.008)、吸烟史(P=0.013)、高血压(P=0.011)、淋巴细胞计数(P=0.028)、纤维蛋白原(P=0.016)及NLR(P<0.001)是AIS患者在接受血管内取栓治疗后预后不良的显著预测因子。构建的预后模型具有76.2%的准确度,78.2%的灵敏度,71.7%的特异度,以及86.7%的阳性预测值。 结论 本研究建立的预测模型有助于临床医师识别接受血管内取栓治疗的AIS高风险患者,并为制定个体化治疗策略提供指导。

     

    Abstract:  Objective To explore the potential prognostic factors of patients with acute ischemic stroke (AIS) after undergoing endovascular mechanical thrombectomy and to construct an effective predictive model. Methods A retrospective analysis of clinical data was conducted on 202 patients with anterior circulation large vessel occlusion AIS from 2 stroke centers. All patients received endovascular mechanical thrombectomy treatment, with treatment and follow-up lasting at least 90 d. Basic demographic characteristics, medical records, and baseline blood biomarker data were collected, and the potential prognostic indicators for AIS after 90 d were screened using least absolute shrinkage and selection operator (LASSO)-logistic regression analysis. Results It was found that alcohol drinking (P=0.029), hypertension (P=0.001), diabetes mellitus (P=0.021), stroke or transient ischemic attack (P=0.049), systolic blood pressure on admission (P=0.009), diastolic blood pressure on admission (P=0.038), blood glucose (P=0.003), white blood cell count (P=0.001), neutrophil count (P=0.001), fibrinogen (P=0.010), systemic immune-inflammation index (P=0.008) and neutrophil-to-lymphocyte ratio (NLR) (P < 0.001) were associated with adverse clinical outcomes. Nine significant prognostic determinants were screened through LASSO-logistic regression analysis. Multivariate logistic regression analysis revealed that male sex (P=0.008), smoking history (P=0.013), hypertension (P=0.011), lymphocyte (P=0.028), fibrinogen (P=0.016), and NLR (P < 0.001) were significant predictive factors for poor prognosis in AIS patients after endovascular thrombectomy treatment. The constructed prognostic model had an accuracy of 76.2%, a sensitivity of 78.2%, a specificity of 71.7%, and a positive predictive value of 86.7%. Conclusion The predictive model established in this study can assist clinicians in identifying high-risk patients with AIS who have undergone endovascular thrombectomy, and it provide guidance for formulating individualized treatment strategies.

     

  • 急性缺血性脑卒中(acute ischemic stroke,AIS)是全球导致长期残疾和死亡率持续上升的主要原因之一[1]。血管内取栓治疗已被证实是治疗大血管闭塞性AIS患者的有效方法,能够改善患者功能并延长治疗时间窗[2]。尽管接受了规范化治疗,仍有相当比例的患者预后不佳,这凸显了建立准确预后预测模型的重要性和临床紧迫性。

    已有多项临床和人口学因素被认定为AIS不良预后的潜在预测因子,包括年龄、基础疾病(如高血压和糖尿病)及脑卒中的初始严重程度[3]。Flack等[4]发现高血压是一个可修改风险因素,且与AIS的发生率和严重程度相关。近年来,炎症生物标志物如白细胞计数和中性粒细胞与淋巴细胞比值(neutrophil-to-lymphocyte ratio,NLR)被认为是预测脑卒中后恢复情况的潜在指标。缺血性脑卒中发病后,中性粒细胞通过活性氧和肽基精氨酸脱氨酶4释放中性粒细胞胞外诱捕网,在急性期破坏血脑屏障并促进血栓形成,加剧神经损伤[5]。选择性耗竭调节性T细胞会减少少突胶质细胞生成、白质修复和卒中后的功能恢复[6]。Kim等[7]和Bi等[8]研究表明NLR升高与较差的临床结局相关。

    目前,关于AIS预后模型的多中心研究仍较少临床尚缺乏针对血管内治疗后AIS患者的特异性预后预测工具。本研究通过多中心回顾性分析,构建了一个整合临床特征和炎症标志物的预测模型。该模型可在取栓术后24 h内快速评估AIS患者90 d不良预后风险,有助于临床医师早期识别高风险患者以优化监护策略,同时可为临床试验的患者分层提供客观依据。

    本研究为探索性分析,回顾性纳入202例接受了血管内机械取栓治疗的前循环大血管闭塞性AIS患者,来自2家综合性卒中中心(复旦大学附属金山医院2023年2月至2024年5月收治的患者,以及上海交通大学医学院附属第九人民医院2022年3月至2024年4月收治的患者)。所有患者的私人数据均已匿名化并严格保密。本研究经上海交通大学医学院附属第九人民医院伦理委员会审核批准(SH9H-2024-T417-1)和复旦大学附属金山医院伦理委员会审核批准(jszxyy202234),并豁免知情同意(因回顾性研究仅使用匿名数据,且所有操作符合《赫尔辛基宣言》要求)。纳入标准:(1)经CT或MRI检查确诊的前循环缺血性脑卒中;(2)年龄为18~90岁;(3)发病到股动脉穿刺时间≤24 h;(4)进行了血管内取栓治疗;(5)随访至少90 d。排除标准:(1)临床和实验室数据不完整;(2)预后数据缺失;(3)合并恶性肿瘤、活动性感染(CRP>50 mg/L)或终末期肝肾功能不全[估计的肾小球滤过率30 mL·min-1·(1.73 m2-1或Child-Pugh C级];(4)预期生存期<3个月。

    收集患者的年龄、性别、身高、腰围等基本人口学变量,吸烟史、饮酒史、高血压、糖尿病、心房颤动、高脂血症等基础疾病(公认的卒中危险因素)信息,入院时的血压和血糖水平,以及白细胞计数、中性粒细胞计数、淋巴细胞计数、血小板计数、血红蛋白、白蛋白、高密度脂蛋白胆固醇(high density lipoprotein-cholesterol,HDL-C)等实验室检查指标。同时计算炎症复合指标NLR、系统性免疫炎症指数(systemic immune-inflammation index,SII)、血小板与淋巴细胞比值(platelet-to-lymphocyte ratio,PLR)和心脏代谢指数(cardiometabolic index,CMI)。SII=血小板计数(/L)×中性粒细胞计数(/L)/淋巴细胞计数(/L);NLR=中性粒细胞计数(/L)/淋巴细胞计数(/L);PLR=血小板计数(/L)/淋巴细胞计数(/L);CMI=腰围(cm)/身高(cm)×甘油三酯(mmol/L)/HDL-C(mmol/L)。

    血管成功再通定义为改良脑梗死溶栓分级为2b或3级[9]。在门诊或电话随访时,基于90 d的改良Rankin量表(modified Rankin scale,mRS)评估患者预后情况,其中mRS得分0~2分为预后良好,>2分为预后不良。

    使用Stata 17.0软件对数据进行分析。使用Shapiro-Wilk检验分析计量资料的正态性,当数据满足正态分布且方差齐以x±s表示,组间比较采用独立样本t检验;当数据达不到正态分布以MIQR)表示,组间比较采用Mann-Whitney U检验。计数资料以例数和百分数表示,组间比较采用Fisher确切概率法或Pearson χ2检验。采用logistic回归分析预后预测因子。为减少因预测因子过多引起的多重共线性,通过计算方差膨胀因子排除存在多重共线性的变量(方差膨胀因子<5),采用最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)正则化方法并运用10折交叉验证选择最佳正则化参数(λ),使用R 4.3.2软件和glmnet包进行LASSO-logistic回归分析。检验水准(α)为0.05。

    共有224例来自2家中心的接受血管内机械取栓治疗的AIS患者,剔除9例临床数据缺失、6例预后数据缺失、7例合并其他系统性疾病患者,最终纳入202例患者。患者的基线临床特征如表 1所示。202例患者的平均年龄为(69.92±12.96)岁,其中男性患者占61.88%(125/202),既往有高血压病史占67.33%(136/202),入院血压中位数为149/80 mmHg(1 mmHg=0.133 kPa),平均血糖水平为(8.17±3.79)mmol/L;血管成功再通率为87.62%(177/202),术后90 d内死亡率为10.40%(21/202)。分析预后良好组与预后不良组患者的临床特征显示,预后不良组的饮酒史、高血压病史、糖尿病史的患者占比及血糖水平、白细胞计数、中性粒细胞计数、D-二聚体水平、纤维蛋白原水平和炎症指数SII、NLR均高于预后良好组(均P<0.05)。

    表  1  血管内机械取栓术后预后良好和预后不良的AIS患者的临床特征分析
    Table  1  Clinical characteristics of AIS patients after thrombectomy in favorable and unfavourable prognosis groups
    Variable All patients N=202 90 d mRS score≤2 N=74 90 d mRS score>2 N=128 P value
    Demographic characteristic
      Age/year, x±s 69.92±12.96 67.42±12.43 70.00±12.38 0.156
      Male, n (%) 125 (61.88) 49 (66.22) 76 (59.38)
      Waist circumference/cm, x±s 94.58±16.12 94.48±16.15 95.74±15.65 0.250
      Body height/cm, x±s 160.15±14.79 162.24±15.23 159.39±14.12 0.135
      BMI/(kg·m-2), x±s 28.8±7.23 28.77±7.25 29.25±6.91 0.240
      Smoking, n (%) 39 (19.31) 9 (12.16) 30 (23.44) 0.077
      Alcohol use, n (%) 25 (12.38) 4 (5.41) 21 (16.41) 0.039
    Past medical history, n (%)
      Hypertension 136 (67.33) 39 (52.70) 97 (75.78) 0.001
      Atrial fibrillation 67 (33.17) 25 (33.78) 42 (32.81) 1.000
      Diabetes mellitus 52 (25.74) 12 (16.22) 40 (31.25) 0.029
      Hyperlipemia 5 (2.48) 1 (1.35) 4 (3.13) 0.654
      Atherosclerosis 17 (8.42) 5 (6.76) 12 (9.38) 0.702
      Stroke or TIA 33 (16.34) 7 (9.46) 26 (20.31) 0.070
      Coronary heart disease 23 (11.39) 9 (12.16) 14 (10.94) 0.973
    Laboratory characteristic
      Blood glucose/(mmol·L-1), x±s 8.17±3.79 7.29±2.88 9.05±4.25 0.002
      White blood cell/(L-1, ×109), x±s 9.39±3.24 8.75±2.59 10.47±3.72 <0.001
      Neutrophil/(L-1, ×109), x±s 7.78±2.88 7.03±2.58 8.62±3.29 <0.001
      Lymphocyte/(L-1, ×109), x±s 1.15±0.49 1.19±0.43 1.13±0.64 0.499
      Platelet/(L-1, ×109), x±s 205.29±69.38 207.88±70.64 202.45±66.27 0.584
      Triglyceride/(mmol·L-1), x±s 2.45±1.69 2.97±1.57 1.32±0.87 0.232
      Total cholesterol/(mmol·L-1), x±s 10.52±1.53 11.16±0.53 9.72±0.97 0.170
      HDL-C/(mmol·L-1), x±s 1.14±0.51 1.16±0.31 1.28±0.90 0.278
      Serum creatinine/(μmol·L-1), x±s 74.29±23.16 72.23±21.07 76.80±25.96 0.199
      APTT/s, x±s 26.79±5.42 26.75±3.70 26.82±6.68 0.928
      D-dimer/(mg·L-1), M (IQR) 1.28 (3.02) 0.89 (1.45) 1.91 (3.84) <0.001
      Fibrinogen/(g·L-1), x±s 2.93±0.87 2.70±0.79 3.10±1.10 0.007
      SII, x±s 1 895.73±1 305.88 1 450.99±982.47 1 927.90±1 275.42 0.006
      Cardiometabolic index, x±s 1.19±0.99 1.12±0.70 1.24±1.06 0.389
      Platelet-to-lymphocyte ratio, x±s 212.73±111.23 197.96±97.31 223.75±122.85 0.124
      Neutrophil-to-lymphocyte ratio, x±s 8.89±4.38 6.74±3.33 9.95±6.30 <0.001
    AIS: Acute ischemic stroke; mRS: Modified Rankin scale; BMI: Body mass index; TIA: Transient ischemic attack; HDL-C: High density lipoprotein-cholesterol; APTT: Activated partial thromboplastin time; SII: Systemic immune-inflammation index.

    单因素logistic回归分析显示,饮酒、高血压、糖尿病和卒中或短暂性脑缺血发作史,以及入院时收缩压和舒张压、血糖、白细胞计数、中性粒细胞计数、纤维蛋白原、SII和NLR均与不良预后相关(均P<0.05,表 2)。

    表  2  AIS患者血管内机械取栓术后90 d预后因素的单因素logistic回归分析
    Table  2  Univariate logistic regression for predicting 90 d outcomes in AIS patients after thrombectomy
    Variable OR (95%CI) P value
    Age 1.02 (0.99, 1.04) 0.157
    Gender 0.75 (0.41, 1.35) 0.335
    Smoking 2.21 (1.02, 5.22) 0.054
    Alcohol use 3.43 (1.24, 12.15) 0.029
    Hypertension 2.81 (1.53, 5.20) 0.001
    Atrial fibrillation 0.96 (0.52, 1.77) 0.888
    Diabetes mellitus 2.35 (1.17, 5.01) 0.021
    Hyperlipemia 2.35 (0.34, 46.54) 0.448
    Atherosclerosis 1.43 (0.51, 4.64) 0.520
    Previous history of stroke or TIA 2.44 (1.05, 6.38) 0.049
    Previous coronary heart disease 0.89 (0.37, 2.24) 0.792
    Admission SBP 1.02 (1.00, 1.03) 0.009
    Admission DBP 1.02 (1.00, 1.05) 0.038
    Blood glucose 1.17 (1.06, 1.31) 0.003
    White blood cell 1.18 (1.08, 1.31) 0.001
    Neutrophil 1.20 (1.08, 1.34) 0.001
    Lymphocyte 0.84 (0.51, 1.39) 0.498
    Platelet 1.00 (0.99, 1.00) 0.583
    Hemoglobin 1.00 (0.99, 1.01) 0.897
    Albumin 0.96 (0.92, 1.01) 0.109
    Triglyceride 1.36 (0.90, 2.19) 0.177
    Total cholesterol 1.00 (0.76, 1.32) 0.994
    HDL-C 1.37 (0.86, 3.06) 0.320
    Serum creatinine 1.01 (1.00, 1.02) 0.200
    APTT 1.00 (0.95, 1.06) 0.928
    D-dimer 1.03 (0.99, 1.11) 0.324
    Fibrinogen 1.61 (1.15, 2.39) 0.010
    SII 1.00 (1.00, 1.00) 0.008
    CMI 1.16 (0.85, 1.68) 0.392
    NLR 1.15 (1.07, 1.24) <0.001
    PLR 1.00 (1.00, 1.00) 0.132
    AIS: Acute ischemic stroke; TIA: Transient ischemic attack; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HDL-C: High density lipoprotein-cholesterol; APTT: Activated partial thromboplastin time; SII: Systemic immune-inflammation index; CMI: Cardiometabolic index; PLR: Platelet-to-lymphocyte ratio; NLR: Neutrophil-to-lymphocyte ratio; OR: Odds ratio; 95%CI: 95% confidence interval.

    采用LASSO-logistic回归模型进行交叉验证,模型筛选出的相应变量数量如图 2A所示。使用LASSO-logistic回归模型构建风险因素分类器,如图 2B所示。采用双向逐步回归法,通过LASSO-logistic回归分析筛选出9个风险因素,包括性别、吸烟史、高血压、卒中或短暂性脑缺血发作史、入院舒张压、淋巴细胞计数、纤维蛋白原、CMI和NLR。

    图  2  LASSO-logistic回归分析(A)与10折交叉验证(B)结果
    Fig.  2  LASSO-logistic regression analysis (A) and 10-fold cross-validation (B) results
    下载: 全尺寸图片

    多因素logistic回归分析显示(表 3),男性患者预后不良的可能性低于女性患者(OR=0.36,95%CI 0.17~0.75,P=0.008),吸烟(OR=3.67,95%CI 1.36~10.79,P=0.013)、高血压(OR=2.55,95%CI 1.25~5.27,P=0.011)、淋巴细胞计数(OR=2.81,95%CI 1.18~7.48,P=0.028)、纤维蛋白原(OR=1.65,95%CI 1.13~2.58,P=0.016)和NLR(OR=1.29,95%CI 1.15~1.46,P<0.001)均是不良预后的预测因素。利用多因素分析中有统计学意义的6个变量建立分类模型,通过混淆矩阵评估该分类模型的性能,结果显示AUC值为0.82,准确度为76.2%,灵敏度为78.2%,特异度为71.7%,阳性预测值为86.7%。

    表  3  AIS患者血管内机械取栓术后90 d预后因素的多因素logistic回归分析
    Table  3  Multivariate logistic regression for predicting 90 d outcomes in AIS patients after thrombectomy
    Variable Adjusted OR (95%CI) P value
    Gender (male vs female) 0.36 (0.17, 0.75) 0.008
    Smoking 3.67 (1.36, 10.79) 0.013
    Hypertension 2.55 (1.25, 5.27) 0.011
    Previous history of stroke or TIA 2.76 (1.04, 8.16) 0.051
    Admission DBP 1.02 (1.00, 1.05) 0.099
    Lymphocyte 2.81 (1.18, 7.48) 0.028
    Fibrinogen 1.65 (1.13, 2.58) 0.016
    CMI 1.42 (0.95, 2.32) 0.124
    NLR 1.29 (1.15, 1.46) <0.001
    AIS: Acute ischemic stroke; TIA: Transient ischemic attack; DBP: Diastolic blood pressure; CMI: Cardiometabolic index; NLR: Neutrophil-to-lymphocyte ratio; OR: Odds ratio; 95%CI: 95% confidence interval.

    本研究分析了202例接受血管内机械取栓的AIS患者的临床数据,并通过采用单因素logistic回归、LASSO回归与交叉验证及多因素logistic回归方法探讨了与治疗后结局相关的风险因素。本研究结果突显了AIS的复杂性,并强调了可能指导临床医师评估患者预后的关键变量。

    本研究中患者的平均年龄为(69.92±12.96)岁,大多数患者为男性(61.88%,125/202),反映了脑卒中发生率中的性别差异,与既往研究结果一致[10]。高血压是最常见的合并症,总体高血压患者占67.33%(136例),这与既往研究中结果相符[11-12]

    单因素logistic回归分析显示多个变量与不良预后显著相关。其中饮酒是一个强有力的预测因子。虽然此前的研究未显示饮酒与AIS的发生和预后有显著关联[13-14],但本研究结果提示需要进一步明确饮酒量对卒中结局的影响。高血压和糖尿病均为AIS患者预后的显著预测因子。糖尿病与卒中患者不良结局的关联已被多项研究确认,这与高血糖对脑血流和组织修复的不利影响有关[15-16]。因此,无论患者所患脑卒中亚型如何,降糖控制及降压控制都非常重要。白细胞计数和中性粒细胞计数是AIS患者强有力的预后预测因子,提示炎症在卒中的重要发病机制,白细胞计数升高与卒中严重程度和恢复不良相关[17-19]。作为系统性炎症标志物,中性粒细胞、SII与NLR也是显著的预后预测因子,该结果进一步突显了免疫反应在AIS结局中的重要性[20],提示靶向血栓性炎症可能有助于改善脑缺血后结局。

    通过LASSO-logistic回归筛选出9个预后预测因子,包括性别、吸烟史、高血压、既往卒中或短暂性脑缺血发作病史、入院舒张压、淋巴细胞计数、纤维蛋白原、CMI和NLR。该方法能有效地识别具有最高预测价值的变量,并通过惩罚信息量较少的变量减少过拟合[21]

    多因素logistic回归分析结果进一步确认了其中6个关键预测因子的显著性。结果显示,性别与预后显著相关,男性比女性具有较低的不良预后率。这一差异可能与性别差异在卒中发病机制、社会背景及支持差异、更高的心房颤动发生率、性激素水平及女性特有的风险因素(如口服避孕药使用、哺乳和不良妊娠结局)有关[10, 22-24]。吸烟史仍然是强有力的预后不良预测因子,突显其对卒中恢复的不利影响。高血压也是一个显著的预后不良预测因子。既往研究表明,卒中或短暂性脑缺血发作病史可能通过黑素瘤缺乏因子2激活及斑块不稳定等机制增加卒中风险[25]。尽管本研究中该因素的P值略高于0.05(P=0.051),但其OR(2.76)仍具有临床警示意义。淋巴细胞计数、纤维蛋白原和NLR是显著的预后预测因子,进一步证明了系统性炎症在脑卒中结局中的重要性。然而,尚不清楚急性升高的系统性炎症标志物是否由中枢神经系统、全身性炎症过程或两者共同引起。尽管如此,急性生物标志物的变化仍然至关重要,因为它们可以帮助指导临床决策[26]。在单因素分析中显示具有统计学意义的入院舒张压,在多因素模型中并未成为独立预测因子。这可能是因为其效应被模型中其他临床因素所解释或掩盖,提示其对预后的影响可能部分通过其他共存的病理生理途径介导。

    本研究使用混淆矩阵评估了模型的性能,结果显示模型的准确度为76.2%,灵敏度为78.2%,特异度为71.7%,阳性预测值为86.7%。这表明该模型具有较高的灵敏度和精确度,在预测AIS患者不良预后方面具有潜在的实用价值。然而,该模型的预测特异度有待提高。这一局限性可能与未纳入关键影像学指标(如梗死核心体积)有关,因此建议临床应用时结合影像学评估以提高预测准确性。与Heo等[27]2019年建立的AIS预后预测模型(logistic回归模型,AUC值=0.849;深度神经网络模型,AUC值=0.888)相比,本模型的区分度(AUC值=0.82)仍需提升。后续研究计划通过扩大样本量、采用机器学习算法优化模型性能,并在独立外部队列中进行验证。

    本研究存在以下局限性。首先,作为回顾性研究,本研究无法排除选择偏倚。其次,2家卒中中心的数据采集时间范围不同,这种时间差异可能引入潜在的偏倚。尽管通过统计方法对中心差异进行了调整,并且验证了2家卒中中心患者基线特征的同质性,但无法完全排除时间差异对临床实践和结局评估的潜在影响。此外,本研究样本量(n=202)虽通过Power分析满足建模需求(1∶14的事件变量比),但未来需通过更大规模前瞻性队列验证模型泛化性。特别值得注意的是,本预测模型的特异度较低(71.7%),可能导致以下临床问题:(1)假阳性率较高,部分实际预后良好的患者可能被误判为高风险,引发不必要的干预;(2)在临床决策支持系统中可能降低医师对模型预测结果的信任度,特别是当预测与临床判断不一致时。为提升模型的临床应用价值,建议未来从以下角度优化研究:(1)扩大样本量并纳入更多预测变量;(2)整合影像学特征与生物标志物,建立多模态预测体系;(3)开发动态预测模型,纳入治疗反应和病情演变参数。例如纳入术后24 h内NLR的动态变化轨迹;探索临床干预阈值(如当NLR>5时启动抗炎治疗),为构建临床决策支持系统奠定基础。总之,虽然本模型对AIS预后预测具有一定价值,但临床应用时需结合临床判断并充分考虑其特异度低的不足。

    本研究构建的预测模型具有显著的临床实用价值。首先,模型仅需常规临床数据和实验室指标,无需特殊设备,便于各级医疗机构推广应用,可辅助临床医师快速识别高危患者。其次,创新性地整合了炎症标志物与传统临床因素,实现了更全面的风险评估,为个体化治疗决策提供了新依据。特别是对于NLR等炎症指标异常升高的患者,提示可能需要加强临床干预。最后,模型的预测结果可指导康复时机的选择:对预后不良高风险患者建议早期强化康复,而对预后良好者则可采取阶梯式功能训练方案,优化康复资源配置。

    本研究为AIS患者接受血管内治疗后的不良预后的临床预测因子提供了有价值的见解,研究结果强调了应关注可被改变的危险因素,如高血压、饮酒及炎症标志物(白细胞计数和NLR)。通过采用LASSO回归和多因素logistic回归识别了一组可靠的预测因子,这些因子可能有助于临床决策并为制定治疗策略提供指导。未来的研究需要进一步验证这些结果并探讨针对这些风险因素的潜在治疗干预措施。

  • 图  2   LASSO-logistic回归分析(A)与10折交叉验证(B)结果

    Fig.  2   LASSO-logistic regression analysis (A) and 10-fold cross-validation (B) results

    下载: 全尺寸图片

    表  1   血管内机械取栓术后预后良好和预后不良的AIS患者的临床特征分析

    Table  1   Clinical characteristics of AIS patients after thrombectomy in favorable and unfavourable prognosis groups

    Variable All patients N=202 90 d mRS score≤2 N=74 90 d mRS score>2 N=128 P value
    Demographic characteristic
      Age/year, x±s 69.92±12.96 67.42±12.43 70.00±12.38 0.156
      Male, n (%) 125 (61.88) 49 (66.22) 76 (59.38)
      Waist circumference/cm, x±s 94.58±16.12 94.48±16.15 95.74±15.65 0.250
      Body height/cm, x±s 160.15±14.79 162.24±15.23 159.39±14.12 0.135
      BMI/(kg·m-2), x±s 28.8±7.23 28.77±7.25 29.25±6.91 0.240
      Smoking, n (%) 39 (19.31) 9 (12.16) 30 (23.44) 0.077
      Alcohol use, n (%) 25 (12.38) 4 (5.41) 21 (16.41) 0.039
    Past medical history, n (%)
      Hypertension 136 (67.33) 39 (52.70) 97 (75.78) 0.001
      Atrial fibrillation 67 (33.17) 25 (33.78) 42 (32.81) 1.000
      Diabetes mellitus 52 (25.74) 12 (16.22) 40 (31.25) 0.029
      Hyperlipemia 5 (2.48) 1 (1.35) 4 (3.13) 0.654
      Atherosclerosis 17 (8.42) 5 (6.76) 12 (9.38) 0.702
      Stroke or TIA 33 (16.34) 7 (9.46) 26 (20.31) 0.070
      Coronary heart disease 23 (11.39) 9 (12.16) 14 (10.94) 0.973
    Laboratory characteristic
      Blood glucose/(mmol·L-1), x±s 8.17±3.79 7.29±2.88 9.05±4.25 0.002
      White blood cell/(L-1, ×109), x±s 9.39±3.24 8.75±2.59 10.47±3.72 <0.001
      Neutrophil/(L-1, ×109), x±s 7.78±2.88 7.03±2.58 8.62±3.29 <0.001
      Lymphocyte/(L-1, ×109), x±s 1.15±0.49 1.19±0.43 1.13±0.64 0.499
      Platelet/(L-1, ×109), x±s 205.29±69.38 207.88±70.64 202.45±66.27 0.584
      Triglyceride/(mmol·L-1), x±s 2.45±1.69 2.97±1.57 1.32±0.87 0.232
      Total cholesterol/(mmol·L-1), x±s 10.52±1.53 11.16±0.53 9.72±0.97 0.170
      HDL-C/(mmol·L-1), x±s 1.14±0.51 1.16±0.31 1.28±0.90 0.278
      Serum creatinine/(μmol·L-1), x±s 74.29±23.16 72.23±21.07 76.80±25.96 0.199
      APTT/s, x±s 26.79±5.42 26.75±3.70 26.82±6.68 0.928
      D-dimer/(mg·L-1), M (IQR) 1.28 (3.02) 0.89 (1.45) 1.91 (3.84) <0.001
      Fibrinogen/(g·L-1), x±s 2.93±0.87 2.70±0.79 3.10±1.10 0.007
      SII, x±s 1 895.73±1 305.88 1 450.99±982.47 1 927.90±1 275.42 0.006
      Cardiometabolic index, x±s 1.19±0.99 1.12±0.70 1.24±1.06 0.389
      Platelet-to-lymphocyte ratio, x±s 212.73±111.23 197.96±97.31 223.75±122.85 0.124
      Neutrophil-to-lymphocyte ratio, x±s 8.89±4.38 6.74±3.33 9.95±6.30 <0.001
    AIS: Acute ischemic stroke; mRS: Modified Rankin scale; BMI: Body mass index; TIA: Transient ischemic attack; HDL-C: High density lipoprotein-cholesterol; APTT: Activated partial thromboplastin time; SII: Systemic immune-inflammation index.

    表  2   AIS患者血管内机械取栓术后90 d预后因素的单因素logistic回归分析

    Table  2   Univariate logistic regression for predicting 90 d outcomes in AIS patients after thrombectomy

    Variable OR (95%CI) P value
    Age 1.02 (0.99, 1.04) 0.157
    Gender 0.75 (0.41, 1.35) 0.335
    Smoking 2.21 (1.02, 5.22) 0.054
    Alcohol use 3.43 (1.24, 12.15) 0.029
    Hypertension 2.81 (1.53, 5.20) 0.001
    Atrial fibrillation 0.96 (0.52, 1.77) 0.888
    Diabetes mellitus 2.35 (1.17, 5.01) 0.021
    Hyperlipemia 2.35 (0.34, 46.54) 0.448
    Atherosclerosis 1.43 (0.51, 4.64) 0.520
    Previous history of stroke or TIA 2.44 (1.05, 6.38) 0.049
    Previous coronary heart disease 0.89 (0.37, 2.24) 0.792
    Admission SBP 1.02 (1.00, 1.03) 0.009
    Admission DBP 1.02 (1.00, 1.05) 0.038
    Blood glucose 1.17 (1.06, 1.31) 0.003
    White blood cell 1.18 (1.08, 1.31) 0.001
    Neutrophil 1.20 (1.08, 1.34) 0.001
    Lymphocyte 0.84 (0.51, 1.39) 0.498
    Platelet 1.00 (0.99, 1.00) 0.583
    Hemoglobin 1.00 (0.99, 1.01) 0.897
    Albumin 0.96 (0.92, 1.01) 0.109
    Triglyceride 1.36 (0.90, 2.19) 0.177
    Total cholesterol 1.00 (0.76, 1.32) 0.994
    HDL-C 1.37 (0.86, 3.06) 0.320
    Serum creatinine 1.01 (1.00, 1.02) 0.200
    APTT 1.00 (0.95, 1.06) 0.928
    D-dimer 1.03 (0.99, 1.11) 0.324
    Fibrinogen 1.61 (1.15, 2.39) 0.010
    SII 1.00 (1.00, 1.00) 0.008
    CMI 1.16 (0.85, 1.68) 0.392
    NLR 1.15 (1.07, 1.24) <0.001
    PLR 1.00 (1.00, 1.00) 0.132
    AIS: Acute ischemic stroke; TIA: Transient ischemic attack; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HDL-C: High density lipoprotein-cholesterol; APTT: Activated partial thromboplastin time; SII: Systemic immune-inflammation index; CMI: Cardiometabolic index; PLR: Platelet-to-lymphocyte ratio; NLR: Neutrophil-to-lymphocyte ratio; OR: Odds ratio; 95%CI: 95% confidence interval.

    表  3   AIS患者血管内机械取栓术后90 d预后因素的多因素logistic回归分析

    Table  3   Multivariate logistic regression for predicting 90 d outcomes in AIS patients after thrombectomy

    Variable Adjusted OR (95%CI) P value
    Gender (male vs female) 0.36 (0.17, 0.75) 0.008
    Smoking 3.67 (1.36, 10.79) 0.013
    Hypertension 2.55 (1.25, 5.27) 0.011
    Previous history of stroke or TIA 2.76 (1.04, 8.16) 0.051
    Admission DBP 1.02 (1.00, 1.05) 0.099
    Lymphocyte 2.81 (1.18, 7.48) 0.028
    Fibrinogen 1.65 (1.13, 2.58) 0.016
    CMI 1.42 (0.95, 2.32) 0.124
    NLR 1.29 (1.15, 1.46) <0.001
    AIS: Acute ischemic stroke; TIA: Transient ischemic attack; DBP: Diastolic blood pressure; CMI: Cardiometabolic index; NLR: Neutrophil-to-lymphocyte ratio; OR: Odds ratio; 95%CI: 95% confidence interval.
  • [1] DING Q, LIU S, YAO Y, et al. Global, regional, and national burden of ischemic stroke, 1990-2019[J]. Neurology, 2022, 98(3): e279-e290. DOI: 10.1212/WNL.0000000000013115.
    [2] NOGUEIRA R G, JADHAV A P, HAUSSEN D C, et al. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct[J]. N Engl J Med, 2018, 378(1): 11-21. DOI: 10.1056/NEJMoa1706442.
    [3] EKKER M S, VERHOEVEN J I, SCHELLEKENS M M I, et al. Risk factors and causes of ischemic stroke in 1322 young adults[J]. Stroke, 2023, 54(2): 439-447. DOI: 10.1161/STROKEAHA.122.040524.
    [4] FLACK J M, ADEKOLA B. Blood pressure and the new ACC/AHA hypertension guidelines[J]. Trends Cardiovasc Med, 2020, 30(3): 160-164. DOI: 10.1016/j.tcm.2019.05.003.
    [5] LI C, XING Y, ZHANG Y, et al. Neutrophil extracellular traps exacerbate ischemic brain damage[J]. Mol Neurobiol, 2022, 59(1): 643-656. DOI: 10.1007/s12035-021-02635-z.
    [6] SHI L, SUN Z, SU W, et al. Treg cell-derived osteopontin promotes microglia-mediated white matter repair after ischemic stroke[J]. Immunity, 2021, 54(7): 1527-1542.e8. DOI: 10.1016/j.immuni.2021.04.022.
    [7] KIM M S, HEO M Y, JOO H J, et al. Neutrophil-to-lymphocyte ratio as a predictor of short-term functional outcomes in acute ischemic stroke patients[J]. Int J Environ Res Public Health, 2023, 20(2): 898. DOI: 10.3390/ijerph20020898.
    [8] BI Y, SHEN J, CHEN S C, et al. Prognostic value of neutrophil to lymphocyte ratio in acute ischemic stroke after reperfusion therapy[J]. Sci Rep, 2021, 11(1): 6177. DOI: 10.1038/s41598-021-85373-5.
    [9] YOO A J, SIMONSEN C Z, PRABHAKARAN S, et al. Refining angiographic biomarkers of revascularization: improving outcome prediction after intra-arterial therapy[J]. Stroke, 2013, 44(9): 2509-2512. DOI: 10.1161/STROKEAHA.113.001990.
    [10] REEVES M J, BUSHNELL C D, HOWARD G, et al. Sex differences in stroke: epidemiology, clinical presentation, medical care, and outcomes[J]. Lancet Neurol, 2008, 7(10): 915-926. DOI: 10.1016/S1474-4422(08)70193-5.
    [11] BARUA L, FARUQUE M, BANIK P C, et al. Agreement between 2017 ACC/AHA hypertension clinical practice guidelines and seventh report of the joint national committee guidelines to estimate prevalence of postmenopausal hypertension in a rural area of Bangladesh: a cross sectional study[J]. Medicina (Kaunas), 2019, 55(7): 315. DOI: 10.3390/medicina55070315.
    [12] ADAMS H P Jr, BENDIXEN B H, KAPPELLE L J, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment[J]. Stroke, 1993, 24(1): 35-41. DOI: 10.1161/01.str.24.1.35.
    [13] XU B, WU Q, LA R, et al. Is systemic inflammation a missing link between cardiometabolic index with mortality?Evidence from a large population-based study[J]. Cardiovasc Diabetol, 2024, 23(1): 212. DOI: 10.1186/s12933-024-02251-w.
    [14] ZHANG Z, WANG M, GILL D, et al. Genetically predicted smoking and alcohol consumption and functional outcome after ischemic stroke[J]. Neurology, 2022, 99(24): e2693-e2698. DOI: 10.1212/WNL.0000000000201291.
    [15] DIENER H C, HANKEY G J. Primary and secondary prevention of ischemic stroke and cerebral hemorrhage JACC focus seminar[J]. J Am Coll Cardiol, 2020, 75(15): 1804-1818. DOI: 10.1016/j.jacc.2019.12.072.
    [16] MAIDA C D, DAIDONE M, PACINELLA G, et al. Diabetes and ischemic stroke: an old and new relationship an overview of the close interaction between these diseases[J]. Int J Mol Sci, 2022, 23(4): 2397. DOI: 10.3390/ijms23042397.
    [17] DELONG J H, OHASHI S N, O'CONNOR K C, et al. Inflammatory responses after ischemic stroke[J]. Semin Immunopathol, 2022, 44(5): 625-648. DOI: 10.1007/s00281-022-00943-7.
    [18] SCHULZ C, MASSBERG S. Inflammaging aggravates stroke pathology[J]. Nat Immunol, 2023, 24(6): 887-888. DOI: 10.1038/s41590-023-01516-y.
    [19] CHATURVEDI S, DE MARCHIS G M. Inflammatory biomarkers and stroke subtype: an important new frontier[J]. Neurology, 2024, 102(2): e208098. DOI: 10.1212/WNL.0000000000208098.
    [20] GONG P, LIU Y, GONG Y, et al. The association of neutrophil to lymphocyte ratio, platelet to lymphocyte ratio, and lymphocyte to monocyte ratio with post-thrombolysis early neurological outcomes in patients with acute ischemic stroke[J]. J Neuroinflammation, 2021, 18(1): 51. DOI: 10.1186/s12974-021-02090-6.
    [21] MAO Y, WENG J, XIE Q, et al. Association between dietary inflammatory index and Stroke in the US population: evidence from NHANES 1999-2018[J]. BMC Public Health, 2024, 24(1): 50. DOI: 10.1186/s12889-023-17556-w.
    [22] REXRODE K M, MADSEN T E, YU A Y X, et al. The impact of sex and gender on stroke[J]. Circ Res, 2022, 130(4): 512-528. DOI: 10.1161/CIRCRESAHA.121.319915.
    [23] NAVEED H, ALMASRI M, KAZANI B, et al. Women and stroke: disparities in clinical presentation, severity, and short- and long-term outcomes[J]. Front Neurol, 2023, 14: 1147858. DOI: 10.3389/fneur.2023.1147858.
    [24] KO D, RAHMAN F, MARTINS M A P, et al. Atrial fibrillation in women: treatment[J]. Nat Rev Cardiol, 2017, 14(2): 113-124. DOI: 10.1038/nrcardio.2016.171.
    [25] CAO J, ROTH S, ZHANG S, et al. DNA-sensing inflammasomes cause recurrent atherosclerotic stroke[J]. Nature, 2024, 633(8029): 433-441. DOI: 10.1038/s41586-024-07803-4.
    [26] VISSER K, KOGGEL M, BLAAUW J, et al. Blood-based biomarkers of inflammation in mild traumatic brain injury: a systematic review[J]. Neurosci Biobehav Rev, 2022, 132: 154-168. DOI: 10.1016/j.neubiorev.2021.11.036.
    [27] HEO J, YOON J G, PARK H, et al. Machine learning-based model for prediction of outcomes in acute stroke[J]. Stroke, 2019, 50(5): 1263-1265. DOI: 10.1161/STROKEAHA.118.024293.
WeChat 点击查看大图
图(1)  /  表(3)
出版历程
  • 收稿日期:  2025-02-16
  • 接受日期:  2025-06-04

目录

    /

    返回文章
    返回