Scholarlink

Background <p>Elderly patients have different risk profiles and stroke features compared to younger individuals. However, there is a paucity of data on acute ischemic stroke (AIS) in very old subjects, and insufficient data has been published about prognostication in this population. This study aimed to design a nomogram for predicting hospital mortality of AIS patients over 80 years old in intensive care units (ICU) and to establish clear risk strata via the gray zone approach.</p> Methods <p>In this retrospective cohort study, the baseline data and in-hospital prognosis of patients with stroke from the Medical Information Mart for Intensive Care-IV database were retrieved. Patient outcomes were dichotomized into survival and non-survival based on 28-day prognosis. Independent predictors identified through univariate and multivariable logistic regression were used to construct the nomogram. Model performance was evaluated in the validation set using AUC, calibration curves, and decision curve analysis (DCA). The gray zone approach categorized patients into three risk groups.</p> Results <p>Initially, 929 eligible patients with AIS were identified, and 224 (24.1%) achieved the endpoint event. The multivariable analysis revealed that Male, White, APS-III, GCS, glucose and Mechanical ventilation were independent predictors for AIS, which were incorporated into a nomogram. The model demonstrated strong performance in the validation set, with an area under the curve (AUC) of 0.814. The gray zone approach effectively stratified patients into three distinct risk categories with significantly different mortality rates: low-risk (Total points ≤ 105), medium-risk (105 &lt; Total points &lt; 129), and high-risk (Total points ≥ 129). This stratification was further validated by significant Kaplan-Meier survival differences (log-rank <i>p</i> &lt; 0.001) and a clear dose-response relationship across risk groups (HR = 2.99 per category, <i>p</i> &lt; 0.001). The current study demonstrated that, within the gray zone, male sex and a higher GCS score were associated with survival.</p> Conclusion <p>In conclusion, the nomogram and its associated risk strata offer a practical tool that directly informs prognosis and clinical decision-making for AIS patients over 80 in the ICU.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development and validation of predicting hospital mortality of acute ischemic stroke patients over 80 years in ICU: a retrospective study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月3日

查看全文
  • Lishan Xu,
  • Yimin Chen,
  • Jamir Pitton Rissardo,
  • Mohammad Mofatteh,
  • Ana Leticia Fornari Caprara,
  • Qibei Dai,
  • Yihua He,
  • Shengli An

摘要

Background

Elderly patients have different risk profiles and stroke features compared to younger individuals. However, there is a paucity of data on acute ischemic stroke (AIS) in very old subjects, and insufficient data has been published about prognostication in this population. This study aimed to design a nomogram for predicting hospital mortality of AIS patients over 80 years old in intensive care units (ICU) and to establish clear risk strata via the gray zone approach.

Methods

In this retrospective cohort study, the baseline data and in-hospital prognosis of patients with stroke from the Medical Information Mart for Intensive Care-IV database were retrieved. Patient outcomes were dichotomized into survival and non-survival based on 28-day prognosis. Independent predictors identified through univariate and multivariable logistic regression were used to construct the nomogram. Model performance was evaluated in the validation set using AUC, calibration curves, and decision curve analysis (DCA). The gray zone approach categorized patients into three risk groups.

Results

Initially, 929 eligible patients with AIS were identified, and 224 (24.1%) achieved the endpoint event. The multivariable analysis revealed that Male, White, APS-III, GCS, glucose and Mechanical ventilation were independent predictors for AIS, which were incorporated into a nomogram. The model demonstrated strong performance in the validation set, with an area under the curve (AUC) of 0.814. The gray zone approach effectively stratified patients into three distinct risk categories with significantly different mortality rates: low-risk (Total points ≤ 105), medium-risk (105 < Total points < 129), and high-risk (Total points ≥ 129). This stratification was further validated by significant Kaplan-Meier survival differences (log-rank p < 0.001) and a clear dose-response relationship across risk groups (HR = 2.99 per category, p < 0.001). The current study demonstrated that, within the gray zone, male sex and a higher GCS score were associated with survival.

Conclusion

In conclusion, the nomogram and its associated risk strata offer a practical tool that directly informs prognosis and clinical decision-making for AIS patients over 80 in the ICU.