Objectives

This study aims at exploring the 3-month outcome predicting ability of delayed neurological improvement and the cause of delayed neurological improvement.

Materials and methods

Early neurological improvement and delayed neurological improvement were calculated to represent the neurological improvements. Good functional outcome was defined as a 90-day modified Rankin Scale score 0–2. We used multivariant logistic regression to explore the influential factors of good functional outcome as well as delayed neurological improvement. We applied net reclassification improvement and integrated discrimination improvement to assess the quantitative improvement of the predictive model.

Results

Early neurological improvement was observed in 50 (23%) patients and delayed neurological improvement exhibited in 67 (30%) patients. Early neurological improvement and delayed neurological improvement were both independent predictive factors to good functional outcome. In the basic model (adjusted for age, admission glucose level, baseline National Institute of Health Stroke Scale, and complications and number of retrieval attempts), early neurological improvement and delayed neurological improvement statistically improved the predictive ability (early neurological improvement: net reclassification improvement = 0.34, 95% confidence interval, 95% confidential interval (0.06, 0.69); integrated discrimination improvement = 0.05, p < 0.001; delayed neurological improvement: net reclassification improvement = 0.79, 95% confidential interval (0.47, 1.12); integrated discrimination improvement = 0.14, p < 0.001) delayed neurological improvement could predict clinical outcomes more accurately than early neurological improvement (early neurological improvement vs. delayed neurological improvement: integrated discrimination improvement = 0.09, p < 0.001). Moreover, delayed neurological improvement was affected by hypertension (odds ratio = 0.40, 95% CI (0.18, 0.88), p = 0.02), early neurological improvement (odds ratio = 20.10, 95% confidential interval (8.24, 19.02), p < 0.001), number of retrieval attempts (odds ratio = 0.39, 95% confidential interval (0.24, 0.66), p < 0.001), and complication (odds ratio = 0.25, 95% confidential interval (0.12, 0.54), p < 0.001).

Conclusions

Delayed neurological improvement could predict clinical outcomes more accurately than early neurological improvement. Hypertension, early neurological improvement, numbers of retrieval attempts, and complications were all predicting factors to delayed neurological improvement.

中文翻译：

---

延迟的神经功能改善可预测血管内血栓切除术患者的长期临床结果

目标

本研究旨在探讨延迟神经功能改善的3个月结果预测能力以及延迟神经功能改善的原因。

材料和方法

计算早期神经功能改善和延迟神经功能改善来代表神经功能改善。良好的功能结果定义为 90 天改良 Rankin 量表评分 0-2。我们使用多变量逻辑回归来探讨良好功能结果以及延迟神经功能改善的影响因素。我们应用净重分类改进和综合判别改进来评估预测模型的定量改进。

结果

50 名 (23%) 患者观察到早期神经功能改善，67 名 (30%) 患者出现延迟神经功能改善。早期神经功能改善和延迟神经功能改善都是良好功能结果的独立预测因素。在基本模型中（根据年龄、入院血糖水平、基线美国国立卫生研究院卒中量表以及并发症和检索尝试次数进行调整），早期神经功能改善和延迟神经功能改善在统计上提高了预测能力（早期神经功能改善：净重分类改善） = 0.34，95% 置信区间，95% 置信区间 (0.06, 0.69)；综合辨别改善 = 0.05，p  < 0.001；延迟神经系统改善：净重分类改善 = 0.79，95% 置信区间 (0.47, 1.12)；综合辨别改善= 0.14，p  < 0.001）延迟神经功能改善可以比早期神经功能改善更准确地预测临床结果（早期神经功能改善与延迟神经功能改善：综合辨别改善= 0.09，p  < 0.001）。此外，延迟神经功能改善受到高血压（比值比 = 0.40，95% CI（0.18，0.88），p  = 0.02）、早期神经功能改善（比值比 = 20.10，95% 置信区间（8.24，19.02），p  < 0.001）、检索尝试次数（优势比 = 0.39，95% 机密区间 (0.24, 0.66)，p  < 0.001）和并发症（优势比 = 0.25，95% 机密区间 (0.12, 0.54)，p  < 0.001） 。

结论

延迟的神经系统改善可以比早期的神经系统改善更准确地预测临床结果。高血压、早期神经系统改善、检索尝试次数和并发症都是延迟神经系统改善的预测因素。