Age-dependent association between BUN/Cr ratio and in-hospital mortality in older patients with atrial fibrillation: a multicenter explainable machine learning study

Q: What are the key findings of this study?

BUN/Cr ratios 20 were associated with higher adjusted in-hospital mortality vs 10-20 in older ICU patients with AF.

Q: What is the clinical evidence from this study?

Study design: Cohort. Population: Critically ill older patients with atrial fibrillation in the ICU (n=19766). Intervention: Blood urea nitrogen-to-creatinine ratio (BUN/Cr >20 and 20 vs 10–20) (OR 1.56, 95% CI 1.39-1.76, p=<0.001).

Q: What does this research mean for the field?

The blood urea nitrogen-to-creatinine ratio exhibits an age-dependent, U-shaped association with in-hospital mortality in older critically ill patients with atrial fibrillation, with both low ( 20) ratios conferring increased mortality risk. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

Q: What question did this study set out to answer?

To examine the association between the blood urea nitrogen-to-creatinine (BUN/Cr) ratio and in-hospital mortality across older age groups with atrial fibrillation and build an explainable machine learning prediction model.

Hongmei Zheng; Yuxin Zhu; Chongzhe Pei; 薛莉; Hongshi Li; Heng Cai

doi:10.1186/s12872-026-07018-7

November 14, 2026BMC Cardiovascular DisordersOpen Access

Age-dependent association between BUN/Cr ratio and in-hospital mortality in older patients with atrial fibrillation: a multicenter explainable machine learning study

View Full Paper

Ask AI

Bookmark

Share

Key result

High BUN/Cr linked to ~56% higher in-hospital mortality in older ICU patients with AF.

OR 1.56
95% CI 1.39-1.76
P<0.001
n=19,766

Why the study?

The prognostic significance of the blood urea nitrogen-to-creatinine ratio across older age groups in critically ill patients with atrial fibrillation remains unclear.

Is the blood urea nitrogen-to-creatinine (BUN/Cr) ratio associated with all-cause in-hospital mortality in critically ill patients aged ≥65 years with atrial fibrillation, and can an explainable machine learning model accurately predict this outcome?

Key Points

To examine the association between the blood urea nitrogen-to-creatinine (BUN/Cr) ratio and in-hospital mortality across older age groups with atrial fibrillation and build an explainable machine learning prediction model.
Retrospective observational study of patients aged ≥ 65 years with atrial fibrillation alive in the intensive care unit (ICU) at 24 hours from the MIMIC-IV database (N=12,887) and an external validation cohort from the eICU database (N=6,879).
Multivariable logistic regression and restricted cubic splines evaluated nonlinear associations and age-related interactions with all-cause in-hospital mortality.
Five machine learning algorithms were trained on a 70% split of MIMIC-IV (n=9,022), tested on a 30% internal validation set (n=3,865), and externally validated in eICU, with model interpretation performed using SHAP.
In MIMIC-IV, 1,892 patients (14.68%) died during hospitalization; compared with BUN/Cr 10–20, adjusted mortality odds were higher for ratios < 10 (OR 1.45, 95% CI 1.11–1.90) and > 20 (OR 1.56, 95% CI 1.39–1.76), showing a U-shaped curve (P for nonlinearity < 0.001).
Age interactions remained significant after Holm correction, with nonlinear associations observed at ages 65–74 and 75–84 years, but not in patients aged ≥ 85 years (P = 0.338).
The random forest model yielded AUCs of 0.802 internally and 0.754 externally (Brier scores 0.103 and 0.119), showing higher internal discrimination than SOFA and APS III (all Holm-adjusted P < 0.001).

Study Design

Type

Cohort (n=19,766)

Multicenter

Yes

Structured PICO

Is the blood urea nitrogen-to-creatinine (BUN/Cr) ratio associated with all-cause in-hospital mortality in critically ill patients aged ≥65 years with atrial fibrillation, and can an explainable machine learning model accurately predict this outcome?

P

Population

19,766 critically ill patients aged ≥65 years with atrial fibrillation across primary (n=12,887; median age 78 years, 42.62% female) and external validation (n=6,879) ICU cohorts, alive at 24 hours after ICU admission and followed for in-hospital mortality.

E

Exposure

Blood urea nitrogen-to-creatinine (BUN/Cr) ratio categories (<10 and >20) measured within the first 24 hours after ICU admission, and a 16-variable explainable machine learning prediction model (random forest) for mortality risk assessment.

C

Comparator

Medium BUN/Cr ratio category (10–20) for epidemiological association analyses; for the machine learning model, comparisons were made against logistic regression, decision tree, XGBoost, LightGBM, and clinical severity score models (SOFA, APS III, and an exploratory CHADS2-like score).

O

Outcome

All-cause in-hospital mortality occurring after the 24-hour ICU admission landmark during the index hospitalization (including deaths after ICU discharge but before hospital discharge).hard clinical

Main Result

Odds Ratio: 1.56 (95% CI 1.39–1.76)

Absolute Event Rate: 17.75% vs 10.67%

p-value: p=<0.001

In older critically ill patients with atrial fibrillation, the BUN/Cr ratio exhibits an age-dependent, nonlinear U-shaped association with in-hospital mortality, and a random forest prediction model incorporating this ratio outperformed traditional ICU severity scores across two large databases.

Limitations

Retrospective observational design
Exclusion of candidate variables (marital status, SBP, DBP, INR, APTT, and anion gap) due to extraction limitations or excessive missingness in the eICU database

Ask about this paper

Answers are grounded in this paper and cite their sources.

Implication

BUN/Cr extremes were associated with higher mortality in older AF ICU patients; leaves open whether the ratio adds value to existing risk models.

Abstract

The prognostic significance of the blood urea nitrogen-to-creatinine ratio (BUN/Cr) across older age groups remains unclear. We investigated its association with in-hospital mortality in older critically ill patients with atrial fibrillation (AF) and developed an explainable machine learning prediction model. This retrospective study included patients aged ≥ 65 years with AF who were alive and remained in the intensive care unit (ICU) at 24 h after admission. Multivariable logistic regression and restricted cubic splines assessed associations between BUN/Cr and all-cause in-hospital mortality in the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Age-related effect modification was evaluated using formal interaction tests with multiple-testing correction. Five machine learning algorithms were developed using a 70% training set and compared in a 30% internal validation set. The random forest model underwent external validation in the eICU Collaborative Research Database (eICU) and interpretation using SHapley Additive exPlanations (SHAP). The MIMIC-IV cohort included 12,887 patients, of whom 1,892 (14.68%) died during hospitalization; the external cohort included 6,879 patients. Compared with BUN/Cr 10–20, ratios < 10 and > 20 were associated with higher adjusted mortality odds: odds ratios (ORs) of 1.45 (95% confidence interval [CI], 1.11–1.90) and 1.56 (1.39–1.76), respectively. The overall association was U-shaped (P for nonlinearity < 0.001). Age interactions remained significant after Holm correction. Nonlinearity was evident at ages 65–74 and 75–84 years, but not ≥ 85 years ( P = 0.338). Random forest achieved areas under the receiver operating characteristic curve (AUCs) of 0.802 internally and 0.754 externally, with Brier scores of 0.103 and 0.119. Its internal discrimination exceeded Sequential Organ Failure Assessment (SOFA), Acute Physiology Score III (APS III), and an exploratory CHADS2-like model (heart failure, hypertension, age ≥ 75 years, diabetes mellitus, and stroke or transient ischemic attack) (all Holm-adjusted P < 0.001). A web-based calculator was developed. BUN/Cr showed a nonlinear association with in-hospital mortality that varied with age. The random forest model demonstrated discrimination across two databases, supporting further evaluation for individualized risk assessment.

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Zheng et al. (2026) conducted a cohort in Critically ill older patients with atrial fibrillation in the ICU (n=19,766). Blood urea nitrogen-to-creatinine ratio (BUN/Cr >20 and <10) vs. BUN/Cr ratio 10–20 was evaluated on All-cause in-hospital mortality (for BUN/Cr >20 vs 10–20) (OR 1.56, 95% CI 1.39-1.76, p=<0.001). BUN/Cr ratios <10 and >20 were associated with higher adjusted in-hospital mortality vs 10-20 in older ICU patients with AF.

synapsesocial.com/papers/6af9ebcdd393cbc380a05052 https://doi.org/10.1186/s12872-026-07018-7

View Full Paper

Ask AI

Bookmark

Share

Age-dependent association between BUN/Cr ratio and in-hospital mortality in older patients with atrial fibrillation: a multicenter explainable machine learning study

Key Points

Study Design

Structured PICO

Main Result

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Age-dependent association between BUN/Cr ratio and in-hospital mortality in older patients with atrial fibrillation: a multicenter explainable machine learning study

Key Points

Study Design

Structured PICO

Main Result

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading