Scholarlink

Background <p>Acute kidney injury (AKI) is a common and life-threatening condition in intensive care unit (ICU) patients. The relationship between body mass index (BMI) and survival outcomes in the Kidney Disease: Improving Global Outcomes (KDIGO) stage 3 AKI remains unclear. Clarifying this association is essential for risk stratification and optimizing individualized management in critically ill populations.</p> Methods <p>A retrospective cohort study was conducted using the Medical Information Mart for Intensive Care IV (MIMIC-IV) 3.0 database, including 5609 ICU patients with KDIGO stage 3 AKI. Patients were categorized according to the World Health Organization (WHO) BMI classification. The primary outcomes were all-cause mortality at 7, 28, and 365 days after ICU admission. Statistical analyses included restricted cubic spline (RCS) modelling, logistic regression, and Kaplan–Meier (KM) survival curves to evaluate the association between BMI and mortality. A random forest-based Boruta algorithm was applied for feature selection, and multiple machine learning models were developed to predict mortality risk, with their performances compared to identify the optimal approach.</p> Results <p>Higher BMI was inversely associated with mortality risk, with the protective effect becoming more pronounced over time. Patients with higher BMI had significantly lower 28-day and 365-day mortality (OR &lt; 1, p &lt; 0.05). RCS analysis confirmed a nonlinear protective relationship between BMI and mortality. However, this survival benefit was attenuated in patients who received continuous renal replacement therapy (CRRT), suggesting an interaction between BMI and CRRT on survival outcomes. Among various predictive models, the Light Gradient Boosting Machine (LightGBM) demonstrated the highest predictive accuracy.</p> Conclusions <p>LightGBM demonstrated the highest predictive performance for mortality risk in ICU patients with KDIGO stage 3 AKI. Higher BMI was independently associated with improved short- and long-term survival, although this protective effect was attenuated in those receiving CRRT. These findings support incorporating BMI into risk stratification and prognostic modelling for KDIGO stage 3 AKI patients, while accounting for the modifying impact of renal replacement therapy.</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

Obesity and mortality in KDIGO stage 3 acute kidney injury: a machine learning-driven retrospective cohort study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月5日

查看全文
  • Qi-Cong Li,
  • Hao-Jie Jin,
  • Tian-Pei Mou,
  • Qun-Li Li,
  • Min-Hao Zhang,
  • Wei-Yi Xia,
  • Zi-Yi Huang,
  • Xi-Hao Zhong,
  • Xiang-Tao Zheng

摘要

Background

Acute kidney injury (AKI) is a common and life-threatening condition in intensive care unit (ICU) patients. The relationship between body mass index (BMI) and survival outcomes in the Kidney Disease: Improving Global Outcomes (KDIGO) stage 3 AKI remains unclear. Clarifying this association is essential for risk stratification and optimizing individualized management in critically ill populations.

Methods

A retrospective cohort study was conducted using the Medical Information Mart for Intensive Care IV (MIMIC-IV) 3.0 database, including 5609 ICU patients with KDIGO stage 3 AKI. Patients were categorized according to the World Health Organization (WHO) BMI classification. The primary outcomes were all-cause mortality at 7, 28, and 365 days after ICU admission. Statistical analyses included restricted cubic spline (RCS) modelling, logistic regression, and Kaplan–Meier (KM) survival curves to evaluate the association between BMI and mortality. A random forest-based Boruta algorithm was applied for feature selection, and multiple machine learning models were developed to predict mortality risk, with their performances compared to identify the optimal approach.

Results

Higher BMI was inversely associated with mortality risk, with the protective effect becoming more pronounced over time. Patients with higher BMI had significantly lower 28-day and 365-day mortality (OR < 1, p < 0.05). RCS analysis confirmed a nonlinear protective relationship between BMI and mortality. However, this survival benefit was attenuated in patients who received continuous renal replacement therapy (CRRT), suggesting an interaction between BMI and CRRT on survival outcomes. Among various predictive models, the Light Gradient Boosting Machine (LightGBM) demonstrated the highest predictive accuracy.

Conclusions

LightGBM demonstrated the highest predictive performance for mortality risk in ICU patients with KDIGO stage 3 AKI. Higher BMI was independently associated with improved short- and long-term survival, although this protective effect was attenuated in those receiving CRRT. These findings support incorporating BMI into risk stratification and prognostic modelling for KDIGO stage 3 AKI patients, while accounting for the modifying impact of renal replacement therapy.