Individualized drug therapy and survival prediction in ICU patients with acute kidney injury: construction and validation of a nomogram

Abstract Background Acute kidney injury (AKI) is defined by a sharp decrease in the estimated glomerular filtration rate (eGFR). However, the impact of medication history on the survival of AKI patients has received little attention. Hence, it is necessary to investigate the potential of medication...

Full description

Saved in:
Bibliographic Details
Main Authors: Rui Yang, Xiaozhe Su, Ziqi Liu, Shuai Shao, Yinhuai Wang, Hao Su, Haiqing He
Format: Article
Language:English
Published: BMC 2025-02-01
Series:European Journal of Medical Research
Subjects:
Online Access:https://doi.org/10.1186/s40001-025-02300-4
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Background Acute kidney injury (AKI) is defined by a sharp decrease in the estimated glomerular filtration rate (eGFR). However, the impact of medication history on the survival of AKI patients has received little attention. Hence, it is necessary to investigate the potential of medication history as a predictor of survival outcomes among AKI patients in the intensive care unit (ICU). Methods Critically ill AKI patients were sourced from the MIMIC-IV database. To ascertain significant, drug-related, independent predictors of survival, univariate Cox analysis and stepwise Cox regression were performed. Based on the identified predictor, a nomogram was developed to estimate the individualized survival probability for AKI patients. Additionally, to address potential confounders among patients with medications referenced in the nomogram, a propensity score matching procedure was applied. Ultimately, a comparative analysis was performed to elucidate the prognostic disparities among these patient subgroups. Results This study enrolled 1,208 patients and developed a nomogram incorporating oxygen flow rate, respiratory frequency, continuous venovenous hemodiafiltration status, age, and medication use (including ibuprofen, epinephrine, cefazolin, warfarin, and vasopressin). The predictive model demonstrated diagnostic accuracy, with AUC values for 1-year, 3-year, and 5-year survival among AKI patients of 0.827, 0.799, and 0.777 in the training dataset, and 0.760, 0.743, and 0.740 in the internal validation dataset, respectively. Kaplan–Meier survival analyses revealed significant differences in survival outcomes among AKI patients based on their exposure to different medications. Conclusions In summary, the developed prediction model demonstrated accuracy for AKI patients in the ICU and helped clinical decision-making. However, future studies will require external validation to confirm these findings.
ISSN:2047-783X