Scholarlink

Background <p>Sepsis is a severe and frequent complication among ischemic stroke patients during hospitalization. The atherogenic index of plasma (AIP), as metabolism-related markers, are closely linked to inflammation. However, their relationship with ischemic stroke sepsis remains unclear.</p> Methods <p>We examined the potential association between sepsis and elevated AIP Levels in 1,531 ischemic stroke intensive care unit patients from the MIMIC-IV database. Logistic regression, restricted cubic splines, and propensity score matching were used to assess associations. Machine learning models were additionally applied to evaluate model performance with and without AIP and to assess its incremental value.</p> Results <p>Higher AIP levels were significantly associated with the development of sepsis. In the fully adjusted model, AIP (OR = 1.75, 95% CI: 1.12–2.71) were independently associated with an increased likelihood of sepsis. Analysis showed that patients in the high AIP group (Q3) exhibited the strongest association with sepsis (OR = 1.51, 95% CI: 1.09–2.09), and this association remained robust after propensity score matching (PSM) (OR = 1.35, 95% CI: 1.01–1.82). In machine learning models, AIP was retained as a contributing feature, and its inclusion improved model discrimination.</p> Conclusion <p>AIP were independently associated with sepsis risk and may complement risk stratification in clinical practice.</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

Association between atherogenic index of plasma and sepsis in critically ill patients with ischemic stroke: a retrospective cohort study using propensity score and machine learning approaches

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月16日

查看全文
  • Fengwei Yao,
  • Lei Liu,
  • Jian Zhang,
  • Xiaolan Chen,
  • Zhijun He

摘要

Background

Sepsis is a severe and frequent complication among ischemic stroke patients during hospitalization. The atherogenic index of plasma (AIP), as metabolism-related markers, are closely linked to inflammation. However, their relationship with ischemic stroke sepsis remains unclear.

Methods

We examined the potential association between sepsis and elevated AIP Levels in 1,531 ischemic stroke intensive care unit patients from the MIMIC-IV database. Logistic regression, restricted cubic splines, and propensity score matching were used to assess associations. Machine learning models were additionally applied to evaluate model performance with and without AIP and to assess its incremental value.

Results

Higher AIP levels were significantly associated with the development of sepsis. In the fully adjusted model, AIP (OR = 1.75, 95% CI: 1.12–2.71) were independently associated with an increased likelihood of sepsis. Analysis showed that patients in the high AIP group (Q3) exhibited the strongest association with sepsis (OR = 1.51, 95% CI: 1.09–2.09), and this association remained robust after propensity score matching (PSM) (OR = 1.35, 95% CI: 1.01–1.82). In machine learning models, AIP was retained as a contributing feature, and its inclusion improved model discrimination.

Conclusion

AIP were independently associated with sepsis risk and may complement risk stratification in clinical practice.