Table 4_PAI-1, MMP-9, and NLR combined with NIHSS for predicting 90-day poor functional outcome in elderly acute ischemic stroke: a prospective observational cohort study.docx

ObjectiveTo investigate whether a multi-marker panel comprising PAI-1, NLR, and MMP-9 enhances prognostication beyond the NIHSS score in elderly patients with acute ischemic stroke, and to develop a clinically applicable nomogram.
MethodsA total of 113 elderly AIS patients and 63 elderly non-AIS controls were prospectively enrolled. Fasting venous blood samples were collected at 06:00 on the first morning after admission (or at 06:00 on the day of admission for overnight admissions), and serum PAI-1, MMP-9, and NLR were measured. Clinical data and NIHSS scores within 24 h of admission were collected. Outcomes were assessed at 90-day follow-up using the modified Rankin Scale (mRS) (favorable outcome: mRS ≤ 2; poor outcome: mRS > 2). Univariate and multivariate logistic regression analyses were performed to identify independent predictors. Three models were constructed: NIHSS alone, biomarkers alone, and their combination. Model performance was evaluated using ROC curves, calibration plots, decision curve analysis (DCA), and bootstrap internal validation.
ResultsSerum levels of PAI-1, MMP-9, and NLR were significantly higher in AIS patients than in controls (all p < 0.01). Among AIS patients, 50 (44.2%) had poor outcomes. In multivariable analysis, NIHSS score (OR, 2.24; 95% CI, 1.60–3.36; p < 0.001), MMP-9 (OR, 1.71; 95% CI, 1.32–2.35; p < 0.001), and NLR (OR, 1.42; 95% CI, 1.12–1.91; p = 0.011) were independently associated with poor outcome; PAI-1 showed a consistent effect direction but did not reach statistical significance (p = 0.077). The combined model achieved an AUC of 0.889 (95% CI, 0.831–0.946), significantly outperforming both the NIHSS-only model (AUC, 0.781; DeLong p = 0.003) and the biomarker-only model (AUC, 0.791; DeLong p = 0.005). The combined model demonstrated excellent calibration (Brier score 0.135), good internal validity (optimism-corrected C-index 0.874), and positive net benefit across a wide threshold probability range (0.15–0.80) on DCA.
ConclusionElevated serum MMP-9 and NLR levels were independently associated with poor short-term prognosis in elderly AIS patients, while PAI-1 showed a consistent direction of association and contributed to the overall performance of the combined model. The nomogram combining these biomarkers with the NIHSS score provides improved risk stratification and may assist early clinical decision-making.

本研究旨在探讨包含纤溶酶原激活物抑制剂-1（plasminogen activator inhibitor-1, PAI-1）、中性粒细胞与淋巴细胞比值（neutrophil-to-lymphocyte ratio, NLR）、基质金属蛋白酶9（matrix metalloproteinase-9, MMP-9）的多标志物组合，是否能在国立卫生研究院卒中量表（National Institutes of Health Stroke Scale, NIHSS）评分之外提升老年急性缺血性卒中（acute ischemic stroke, AIS）患者的预后预测效能，并开发一款临床可用的列线图（nomogram）。
方法：本研究前瞻性纳入113例老年急性缺血性卒中（AIS）患者与63例老年非AIS对照者。于入院后首个清晨6:00采集空腹静脉血标本（过夜入院者于入院当日6:00采集），检测血清PAI-1、MMP-9及NLR水平。收集患者入院24小时内的临床资料与NIHSS评分。于90天随访时采用改良Rankin量表（modified Rankin Scale, mRS）评估预后结局（预后良好定义为mRS≤2分，预后不良定义为mRS>2分）。采用单因素及多因素logistic回归分析筛选独立预测因子，构建三类模型：单纯NIHSS模型、单纯生物标志物模型以及二者联合模型。通过受试者工作特征曲线（receiver operating characteristic curve, ROC曲线）、校准曲线、决策曲线分析（decision curve analysis, DCA）以及Bootstrap内部验证评估模型性能。
结果：AIS患者的血清PAI-1、MMP-9及NLR水平均显著高于对照组（均p<0.01）。在AIS患者中，50例（44.2%）出现预后不良。多因素分析显示，NIHSS评分（优势比OR=2.24，95%置信区间CI：1.60~3.36，p<0.001）、MMP-9（OR=1.71，95%CI：1.32~2.35，p<0.001）及NLR（OR=1.42，95%CI：1.12~1.91，p=0.011）均与预后不良独立相关；PAI-1虽呈现一致的效应方向，但未达到统计学显著性（p=0.077）。联合模型的受试者工作特征曲线下面积（AUC）为0.889（95%CI：0.831~0.946），显著优于单纯NIHSS模型（AUC=0.781，DeLong检验p=0.003）与单纯生物标志物模型（AUC=0.791，DeLong检验p=0.005）。联合模型展现出优异的校准性能（Brier评分0.135）、良好的内部有效性（校正乐观偏倚C指数0.874），且在决策曲线分析（DCA）中覆盖0.15~0.80的宽阈值概率范围时均呈现正向净获益。
结论：老年AIS患者的血清MMP-9与NLR水平升高与短期不良预后独立相关，PAI-1虽呈现一致的关联方向，但其提升了联合模型的整体性能。将上述生物标志物与NIHSS评分相结合的列线图可优化风险分层，有助于辅助早期临床决策制定。